article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
approximately 1% of stars host hot jupiters , ousted from their birthplaces to short - period orbits @xcite via mechanisms that remain debated .
proposed theories fall into two classes : smooth disk migration ( e.g. @xcite ) , and migration via gravitational perturbations , either by stars ( e.g. stellar binary kozai , @xcite ) or sibling planets ( including planetary kozai , e.g. @xcite ; scattering , e.g. @xcite ; and secular chaos , e.g. @xcite ) .
( see @xcite , dmj13 hereafter , for additional references . )
we consider the latter class as also encompassing gravitational perturbations proceeded by disk migration ( e.g. @xcite ) .
migration processes must not only produce hot jupiters heavily studied , extensively observed gas giants orbiting within 0.1 au of their host stars but also populate the region from 0.1 to 1 au , interior to the ice line ( beyond which large , rocky cores can grow and accrete ) but outside the reach of tidal damping forces exerted by the host star .
we call this semi - major axis range the `` valley , '' because it roughly corresponds to the period valley " ( e.g. @xcite ) , the observed dip in the giant planet orbital period ( @xmath1 ) distribution from roughly @xmath2 days . the valley houses gas giants both on highly eccentric and nearly circular orbits .
gas disk migration is unlikely to excite large eccentricities ( e.g. @xcite ) whereas dynamical interactions are unlikely to produce a substantial population of circular orbits .
therefore this eccentricity distribution may point toward intermixing between two different migration mechanisms , one gentle and one violent .
another orbital feature the bimodal distribution of spin - orbit alignments among hot jupiters is sometimes interpreted as evidence for two migration mechanisms @xcite .
however , it may result from stellar torques on the proto - planetary disk @xcite , gravity waves that misalign the star s spin axis @xcite , or two regimes for tidal realignment @xcite .
because tides are negligible in the valley ( except at the most extreme periastron , e.g. hd-80606-b , hd-17156-b ) , we can interpret trends more easily .
excited inclinations and eccentricities can not have been erased by tidal damping .
if _ two _ common mechanisms indeed deliver close - in giant planets , physical properties of the proto - planetary environment may determine which is triggered .
a decade ago , @xcite discovered that giant planets more commonly orbit metal - rich stars , supporting the core accretion formation theory .
independent and follow - up studies confirmed this trend for giant planets ( e.g. @xcite , @xcite @xcite , @xcite , @xcite ) but not small planets @xcite . neither the @xcite nor the @xcite samples exhibited correlations between stellar metallicity and planetary period or eccentricity , but now the radial - velocity ( rv ) sample has quadrupled .
it is time to revisit the planet - metallicity correlation , but now to gain insight into the dynamical evolution of planetary systems _ following _ planet formation .
another motivation is the puzzlingly low occurrence rate of hot jupiters in the vs. rv sample @xcite .
targets have systematically lower metallicities than rv targets .
we will show that differences in the planetary period distribution not just the overall occurrence rate between metal - rich and metal - poor stars may account for the discrepancy .
we uncover new stellar metallicity trends in the eccentricities of giant valley planets ( 2 ) , eccentricities of giant planets tidally circularizing ( 3 ) , and giant planet period distribution ( 4 ) .
these correlations point toward planet - planet interactions as one of two mechanisms for delivering close - in gas giants ( 5 ) .
valley gas giants are unlikely to have formed in situ @xcite and exhibit a range of eccentricities ( @xmath3 ) ( figure 1 ) . here
we consider rv - detected giant planets with @xmath4 , ( queried from the exoplanet orbit database ) , hd-81688-b ( @xmath5 ) , and xi - aql - b ( @xmath6 ) using @xcite s data ; adopt @xcite s @xmath7 for hd-96063-b ; and remove hd-104067-b because the rvs are unavailable . ]
[ eod ] , @xcite ) .
we impose stellar cuts of 4500k@xmath8k and @xmath9@xmath10 throughout the paper to limit our sample to main - sequence stars with with reliable effective temperatures and surface gravities in the input catalogue @xcite .
( we use the sample in 5 . ) under the two migration mechanisms hypothesis , valley planets on nearly circular orbits moved in smoothly through the gaseous proto - planetary disk , whereas those on eccentric orbits were displaced through multi - body interactions . in figure 1 , we emphasize planets with large eccentricities by plotting @xmath11 . this quantity is related to the specific orbital angular momentum , @xmath12 , an important parameter for dynamical interactions .
this scale also minimizes eccentricity bias .
for example , as a result of noise and eccentricity bias , a planet truly on a circular orbit could have a measured @xmath13 .
however , on this scale , @xmath14 would be nearly indistinguishable from @xmath15 .
we divide the sample into planets orbiting metal - rich stars ( [ fe / h]@xmath160 , blue circles ) vs. metal - poor stars ( [ fe / h]@xmath00 , red squares ) . only the metal - rich stars host valley planets with large eccentricities .
the eccentricities of these 68 planets extend up to 0.93 .
in contrast , the 26 valley planets orbiting metal - poor stars are confined to low eccentricities ( @xmath17 ) .
overall , 26% of valley planets orbiting metal - rich stars have eccentricities exceeding that of the most eccentric one orbiting a metal - poor star .
we assess the statistical significance of the low eccentricities of valley planets orbiting metal - poor stars .
we perform a kolmogorov - smirnov
( k - s ) test on the null hypothesis that the eccentricities of the metal - rich and metal - poor sample are drawn from the same distribution .
we reject the null hypothesis with 99.4% confidence .
using a test more sensitive to the tails of distributions , anderson - darling ( a - d ) , we reject the null hypothesis with 99.8% confidence .
the results are insensitive to the exact metallicity cut and statistically significant for any cut located between -0.1 and 0.1 dex .
finally , we perform a monte carlo simulation , randomly selecting 26 planets from the total sample of 94 valley giants . in only 0.4% of @xmath18 trials
is the maximum eccentricity of the 26 planets less than or equal to the observed @xmath19 .
therefore , with 99.6% confidence , we reject the hypothesis that the confinement to low eccentricities of the planets orbiting metal - poor stars results from chance .
although the exact statistical significance is somewhat sensitive to the definition of the valley , which defines the sample size e.g. , if we cut off the valley at 0.8 au instead of 1 au , our confidence is 98.3% it is evident in figure 1 that the trend occurs throughout the valley . as suggested by @xcite in the context of the mutual inclinations of multi - planet systems
, one might expect a threshold metallicity to trigger instability . decreasing planets semi - major axes @xmath20 via gravitational perturbations requires interactions between at least two ( and probably more ) closely - spaced giant planets .
it may be that only metal - rich proto - planetary environments can form such systems .
in contrast , planets on circular orbits would have arrived via disk migration , which can occur regardless of metallicity .
we note that beyond 1 au , the metal - rich and metal - poor sample have similar eccentricity distributions .
planets with @xmath21 au have not necessarily changed their semi - major axes : they may have formed where we observe them .
these planets on eccentric orbits near their formation location may have exchanged angular momentum with another planet or star without requiring the abundance of closely - packed giant planets necessary to drastically alter @xmath22 .
we turn to planets experiencing significant tidal dissipation , detected fixed at 0 in eod fits .
we remove those with poorly - constrained eccentricities : corot-7-b , hat - p-9-b , ogle - tr-10-b , ogle - tr-111-b , tres-1-b , tres-4-b , wasp-13-b , wasp-39-b , wasp-58-b , xo-1-b , xo-5-b .
we include planets whose eccentricities are constrained to be small ( @xmath23 ) , by our fits ( corot-13-b , corot-17-b , wasp-16-b ) or the literature ( corot-7-b , hat - p-1-b , hat - p-4-b , hat - p-8-b , hat - p-12-b , hat - p-27-b , hat - p-39-b , ogle - tr-211-b , kelt-2-ab , wasp-7 , wasp-11-b , wasp-15-b , wasp-21-b , wasp-25-b , wasp-31-b , wasp-35-b , wasp-37-b , wasp-41-b , wasp-42-b , wasp-47-b , wasp-61-b , wasp-62-b , wasp-63-b , wasp-67-b ) .
see the eod for each planet s orbital reference .
] by non - transit surveys ( figure 2 ) and followed up with rv measurements .
we use the stellar and planetary cuts described in 1 .
@xcite and dmj13 used this sample to calculate the abundance of moderately - eccentric proto - hot jupiters . advantageously for this sample , transit surveys are less inclined to target metal - rich stars , yielding planets orbiting metal - poor stars for comparison . to be consistent with @xcite and dmj13 and to avoid eccentricity bias , we classify planets with @xmath24 as eccentric .
the striped region contains planets undergoing tidal circularization along tracks of constant angular momentum ( see @xcite , dmj13 ) to final orbital periods @xmath25 between 3 and 10 days .
( the traditional boundary for hot jupiters is 10 days , and 3 days is the limit above which we still see eccentric giant planets . those with @xmath26 days have much faster tidal circularization rates . )
we only observe eccentric planets orbiting metal - rich stars ( blue circles ) .
we suggest that only giant planets forming in metal - rich systems with multiple giant planets are likely to be scattered onto eccentric orbits that bring them close enough to the star to undergo tidal circularization ( e.g. @xcite ) . to assess the statistical significance
, we perform a monte carlo simulation .
we randomly select 8 host star metallicities from the 53 total host stars in the striped region . in
only 2.3% of 10@xmath27 trials was [ fe / h]@xmath28 for all eight .
we also perform a k - s ( a - d ) test , rejecting with 99.1% ( 99.2% ) confidence the null hypothesis that the host star metallicities of planets in the striped region with @xmath290.2 are drawn from the same distribution as those with @xmath30 0.2 .
the gray error bars are from uncertainties in @xmath31 , not the poisson uncertainties of each individual bin .
the two distributions are consistent at long periods , but the sample lacks a short period pile - up . ] @xmath160 ( middle ) , and with [ fe / h]@xmath00 ( bottom ) . in the metal - rich sample ( middle )
, we recover the shape of the short - period pile - up seen in the rv sample ( black - dashed line , figure 3 ) .
in contrast , the metal - poor sample ( bottom ) is depleted in short - period giants . ]
@xcite found a surprisingly low hot jupiter occurrence rate ( @xmath32 ) the expected number of giant planets per star with @xmath33 days compared to rv surveys ( @xmath34 ) , a trend confirmed by @xcite and @xcite ; all suggested that the systematically lower metallicities of host stars may contribute to the discrepancy . in figure 3
, we compare the period distribution of transiting giant planet candidates detected by the survey ( @xcite ; see also @xcite and @xcite ) applying a radius cut of 8@xmath35 to that expected from the rv sample , the rv sample is not uniform ; we plot it for qualitative comparison .
the expected distribution derived from the period distribution reported by @xcite appears similar .
we therefore interpret the short - period pile - up as real , not due to preferential detection .
for the quantitative calculations in this section , we use the uniform @xcite sample .
] using a normalization constant @xmath31 ( defined below ) .
the two distributions appear consistent beyond 10 days but differ strikingly at short orbital periods : the period distribution lacks a short - period pile - up ( in fact , the absolute giant planet occurrence declines toward short orbital periods , as modeled by @xcite and @xcite ) .
although input catalog ( kic ) metallicity estimates are known to be uncertain @xcite , we can roughly divide the sample into metal - rich ( [ fe / h]@xmath160 ) and metal - poor ( [ fe / h]@xmath00 ) . in figure 4
, we compare the period distributions for giant planets orbiting metal - rich vs. metal - poor stars .
when we limit the sample to [ fe / h]@xmath36 ( row 2 ) , we recover the missing short - period pile - up , which the metal - poor sample ( row 3 ) lacks . performing a k - s test
, we reject with 99.94% confidence the hypothesis that the metal - rich sample and metal - poor sample are drawn from the same distribution .
the results are insensitive to the exact metallicity cut .
we compare the metal - rich(poor ) sample to the rv metal - rich(poor ) sample in figure 5 . in figures 3 and 5
, we compare the observed number of transiting giant planets ( red striped ) to the number expected ( black dashed ) based on the rv sample , @xmath37 where @xmath38 is the observed number of rv planets per bin and @xmath39 is the transit probability .
we set the normalization constant , @xmath31 , using the values ( computed below ) of @xmath32 and @xmath34 : @xmath40 each error bar is due to the uncertainty in @xmath41 . to compute @xmath32
, we follow @xcite , using our own stellar and planetary cuts and the latest sample of candidates @xcite .
the barbara a. mikulski archive for space telescopes ( mast ) supplied the stellar parameters and the nexsci exoplanet archive the transit shape parameters ( duration , depth , @xmath42 , @xmath43 ) .
we obtain @xmath44 for giant planets with @xmath4510 days ( consistent with @xcite and @xcite ) , 1.15@xmath46 for the metal - rich sample , and 0.25@xmath47 for the metal - poor sample .
to compute @xmath34 , we use the stellar and planetary sample from the iconic planet - metallicity correlation @xcite and associated stellar parameters @xcite , the last rv target list to be publicly released .
we obtain @xmath48% for giant planets with @xmath4510 days ( in agreement with @xcite ) , 2.11@xmath49% for those orbiting stars with [
fe / h]@xmath160 , and 0.26@xmath50 for [ fe / h]@xmath00 . with no metallicity cut
, @xmath32 is inconsistent with @xmath34 at the @xmath51 level . in the metal - rich comparison ( figure 5 , left ) ,
we see greater consistency between the and rv distribution than in the full sample ( figure 3 ) .
the metal - rich sample exhibits a short - period pile - up ; the discrepancy between @xmath32 vs. @xmath34 is now only @xmath52 , with the greatest discrepancy in the 3 - 5 day bin .
this improvement motivates a detailed follow - up analysis , including a more precise estimate of @xmath34 using the latest rv target lists .
if follow - up studies find a significant discrepancy between the metal - rich and radial velocity samples , it could be due to the kic metallicity estimates .
using spectroscopic metallicity measurements by @xcite , we find that high kic metallicities do correspond to high spectroscopic metallicities ( with a scatter of about 0.2 dex ) , but the spectroscopic metallicities are systematically higher by 0.1 dex , as also noted by @xcite .
however , we attribute the systematic offset to the fact that stars targeted for spectroscopic follow - up are bright , main - sequence stars in our solar neighborhood and thus have systematically higher metallicities ; in contrast , the kic metallicities were computed assuming a low - metallicity prior , due to the targets being above the galactic plane .
the planetary radius cut may also contribute to the discrepancy .
the @xmath53 cut for the sample corresponds to the rv cut of @xmath54@xmath55 for a planet made of pure hydrogen at a low effective temperature ( e.g. @xcite ) .
however , close - in , low - mass planets may be inflated to @xmath56 and may have a different period distribution , contaminating the sample . in the metal - poor comparison ( figure 5 ,
right ) , the and rv distributions do not appear inconsistent , but it is difficult to judge given the very small sample of rv - detected planets orbiting metal - poor stars .
we found three ways in which the properties of hot jupiters and valley giants depend on host star metallicity : 1 .
gas giants with @xmath57au orbiting metal - rich stars have a range of eccentricities , whereas those orbiting metal - poor stars are restricted to lower eccentricities . 2 .
only metal - rich stars host eccentric proto - hot jupiters undergoing tidal circularization .
the pile - up of short - period giant planets , missing in the sample , is a feature of metal - rich stars and is largely recovered for giants orbiting metal - rich host stars .
hot jupiters and valley giants are both thought to have been displaced from their birthplaces .
therefore these metallicity trends can be understood if smooth disk migration and planet - planet scattering both contribute to the early evolution of systems of giant planets .
we expect disk migration could occur in any system , but only systems packed with giant planets which most easily form around metal - rich stars can scatter giant planets inward to large eccentricities ( trend 1 ) , some of which tides shrink and circularize ( trend 2 ) , creating a pile - up of short - period giants ( trend 3 )
. moreover , these trends support planet - planet interactions ( e.g. scattering , secular chaos , or kozai ) as the dynamical migration mechanism of close - in giant planets , rather than stellar kozai .
this is consistent with previous work by dmj13 arguing that stellar kozai does not produce most hot jupiters , based on the lack of super - eccentric proto - hot jupiters .
we would not expect planet - planet scattering to typically result in nearby companions to hot jupiters , which have been ruled out in the sample by @xcite .
( see also @xcite . )
one possible challenge for our interpretation is the lack of apparent correlation between spin - orbit misalignment and metallicity .
however , spin - orbit misalignments are not necessarily caused by dynamical perturbations , and their interpretation is complicated because measurements have primarily been performed for close - in planets subject to tidal realignment .
we recommend spin - orbit alignment measurements , via spectroscopy @xcite or photometry @xcite , of candidates in the valley , which are typically too distant to be tidally realigned . to support or rule - out the interpretation that these metallicity trends are signatures of planet - planet interactions , we further recommend : 1 ) theoretical assessments of whether planet - planet interaction mechanisms designed to account for hot jupiters can simultaneously produce the observed population of eccentric valley planets , and 2 ) more sophisticated assessments of the trends we report here , using the target lists of recent rv surveys and , as undertaken by @xcite , a careful treatment of false - positives and detection thresholds .
the nsf - grfp ( dge-1144152 ) supports r.i.d .
we thank the referee for the helpful , timely report . our gratitude to john johnson for many illuminating discussions about the three - day pile - up , the period distribution of giant planets , observational approaches to distinguishing the origins of hot jupiters , valuable insights on previous collaborations connected to this investigation , and extensive comments .
we thank daniel fabrycky for many helpful comments , courtney dressing for occurrence rate insights , and subo dong , zachory berta , david charbonneau , sean andrews , matthew holman , jason wright , and kevin schlaufman for useful discussions .
we used the exoplanet orbit database and exoplanet data explorer at exoplanets.org , and the nasa exoplanet archive , operated by caltech , under contract with the nasa exoplanet exploration program .
we include data collected by the mission , funded by the nasa science mission directorate , and are grateful to the team for long and extensive efforts in producing this rich dataset .
some data were obtained from mast .
the association of universities for research in astronomy , inc
. operates stsci under nasa contract nas5 - 26555 .
the nasa office of space science supports mast for non - hst data via grants ( including nnx09af08 g ) and contracts . | gas giants orbiting interior to the ice line are thought to have been displaced from their formation locations by processes that remain debated . here
we uncover several new metallicity trends , which together may indicate that two competing mechanisms deliver close - in giant planets : gentle disk migration , operating in environments with a range of metallicities , and violent planet - planet gravitational interactions , primarily triggered in metal - rich systems in which multiple giant planets can form .
first , we show with 99.6% confidence that giant planets with semi - major axes between 0.1 and 1 au orbiting metal - poor stars ( [ fe / h]@xmath00 ) are confined to lower eccentricities than those orbiting metal - rich stars .
second , we show with 98.3% confidence that eccentric proto - hot jupiters undergoing tidal circularization primarily orbit metal - rich stars . finally , we show that only metal - rich stars host a pile - up of hot jupiters , helping account for the lack of such a pile - up in the overall sample .
migration caused by stellar perturbers ( e.g. stellar kozai ) is unlikely to account for the trends .
these trends further motivate follow - up theoretical work addressing which hot jupiter migration theories can also produce the observed population of eccentric giant planets between 0.1 and 1 au . |
electron anti - neutrinos from large commercial nuclear reactors are playing an important role in the exploration of neutrino oscillations @xcite .
the choice of distance between source and detector allows one to conveniently tune an experiment s sensitivity to either the atmospheric or solar neutrino mass splitting .
kamland @xcite has established a connection between the msw effect in the sun @xcite and vacuum oscillations with anti - neutrinos , and has provided the most accurate measurement of @xmath2 to date .
earlier , chooz @xcite and palo verde @xcite provided what are still the best upper limits on the mixing angle @xmath3 .
a variety of new experiments are being constructed to perform more sensitive measurements of the mixing angle @xmath3 @xcite .
anti - neutrino detectors at very short distance are being explored with the purpose of detecting coherent neutrino - nucleon scattering ( c@xmath4ns ) @xcite , and for testing plutonium diversion at commercial reactors @xcite .
a proper understanding of the systematic effects involved in the modeling of the reactor anti - neutrino flux and energy spectrum is essential for these experiments .
moreover , as most experiments measure @xmath0s from more than one reactor , it is important to understand not only the magnitude of these effects , but also the correlations between uncertainties from different sources . except for the case of c@xmath4ns , which will not be addressed further here , reactor anti -
neutrinos are usually detected using the inverse-@xmath5 reaction @xmath6 whose correlated signature helps reduce backgrounds but is only sensitive to @xmath1s with energy above 1.8 mev .
the kinetic energy of the positron gives a measure of the incoming @xmath1 energy .
neutrino oscillation parameters may be extracted from the data by fitting the observed @xmath0 spectrum @xmath7 to the following equation : @xmath8 here , @xmath9 is the number of target protons , @xmath10 is the energy - dependent detection efficiency , @xmath11 is the detection cross - section , @xmath12 is the oscillation survival probability for @xmath1s traveling a distance @xmath13 from reactor @xmath14 to the detector , and @xmath15 is the @xmath1 spectrum emitted by reactor @xmath14 .
the time variation of the flux may be included in the fit to discriminate the time - varying reactor signal from constant backgrounds .
equation [ eq : nuebarspeccalc ] should technically be multiplied by the detector resolution function and integrated over @xmath16 , but this detail is inconsequential for our analysis .
nuclear power reactors operate on the principle that the fission of u and pu isotopes and the subsequent decays of their daughter fragments release energy , generating heat .
large q - value @xmath5-decays of unstable fission fragments are primarily responsible for the @xmath1 emission of nuclear reactors .
decays of long - lived isotopes in the nuclear fuel and in spent fuel elements stored at the reactor site contribute at the sub - percent level to the detected @xmath1 flux ; this contribution has been treated elsewhere @xcite and is beyond the scope of this paper .
considering only fission reactions , the final term in equation [ eq : nuebarspeccalc ] may be expanded as @xmath17 where @xmath18 is the @xmath1 emission spectrum per fission of isotope @xmath19 , and @xmath20 is the number of fissions of isotope @xmath19 during the data taking period . since @xmath2199.9% of the energy- and @xmath1-producing fissions are from @xmath22u , @xmath23u , @xmath24pu , and @xmath25pu @xcite ( see figure [ fig : pv_fission_rates ] ) , the summation in equation [ eq : reactoremissionspec ] is performed over just these 4 isotopes . for @xmath22u , @xmath24pu , and @xmath25pu ,
the emitted @xmath1 spectra @xmath26 are derived from @xmath5-spectrum measurements of the fissioning of the isotopes by thermal neutrons @xcite .
the uncertainty in the @xmath1 spectral shape derived from these measurements was investigated recently in @xcite . for @xmath23u ,
no measurements are available , so theoretical calculations of its @xmath1 emission must be used @xcite . the @xmath20 can be obtained from detailed simulations of the reactor core throughout the data taking period .
the output @xmath20 must obey the thermal energy constraint @xmath27 where @xmath28 is the total thermal energy produced by the reactor during the time period considered , and @xmath29 denotes the energy released per fission of isotope @xmath19 . to reduce sensitivity to errors in the simulation codes ,
the codes are typically used to obtain not the @xmath20 directly but the fission fractions @xmath30 , where @xmath31 .
measurements of the total generated thermal power taken regularly during reactor operation may then be used to obtain the total numbers of fissions of each isotope using equation [ eq : powersum ] .
the appropriate simulation codes are cumbersome to run , and are often proprietary , and therefore are not always available for direct use by scientific collaborations . obtaining the @xmath32 therefore requires special agreements with reactor operators . depending on the terms of the agreement , fine grained time - dependent fission fractions may be obtained .
fine - grained data for @xmath28 are commonly available , since such information must be made available to nuclear regulatory bodies .
reactor @xmath1 experiments then use this information to calculate the @xmath1 signal using an equation of the following form : @xmath33 where for simplicity we have assumed a single reactor , and have integrated over @xmath16 .
the terms in equations [ eq : nuebarspeccalc ] and [ eq : reactoremissionspec ] not appearing in equation [ eq : nuebarspeccalcmod ] have been grouped into the @xmath34 , the number of detected @xmath1 per fission of isotope @xmath19 .
most recent reactor @xmath1 experiments , including kamland @xcite , chooz @xcite , palo verde @xcite and bugey @xcite , have used equation [ eq : nuebarspeccalcmod ] , or variations on it , to perform their signal calculations . in the first two experiments , the systematic uncertainty in the signal calculation
was taken to be the quadratic sum of the uncertainty in the reactor power measurements which yield @xmath28 and some estimate of the error due to the uncertainty in the fission fractions @xmath32 . in the case of palo verde
this latter component was estimated from comparisons between calculated and measured isotopic concentrations in spent fuel elements @xcite . in the case of kamland
this component was taken from comparisons of a simplified reactor model to detailed simulations @xcite , which assumes that the uncertainty of the simplified reactor model significantly exceeds that of the detailed simulations .
however , no systematic uncertainty was assigned to the detailed reactor simulations . the chooz experiment @xcite leveraged the short baseline bugey observation to limit the reaction cross section uncertainty , scaling their result to agree with the bugey - measured cross section per fission .
this treatment allowed chooz to significantly reduce the error associated with the @xmath1 emission spectra ( @xmath35 ) .
however , their method is legitimate only to the extent that the reactor simulations employed by the chooz and bugey reactor operators reliably model the different reactors with their specific fuel compositions and operation histories . in all cases ,
the combined systematic error was estimated to be within 2 - 3% . at this level ,
the different methods employed by each experiment are likely not overly aggressive . with detector - based uncertainties in current and next - generation reactor
@xmath1 experiments approaching the @xmath361% scale , a more careful treatment of these reactor - specific uncertainties is necessary , particularly for experiments with multiple reactor sources in which correlations become important . in this paper
we outline such a detailed treatment of these uncertainties and demonstrate its application to a counting analysis of @xmath1 s from single and multiple commercial reactor sources . in section [ power_ch ] we discuss in detail the uncertainty in reactor thermal power measurements . in section [ compo_ch ] we examine the monte - carlo estimation of the @xmath20 and , expanding on the treatment in @xcite and @xcite , derive uncertainties for these based on a large body of spent fuel isotopic concentration comparisons used to verify the codes . in section [ analysis_ch ] we combine the uncertainties from @xmath28 and the @xmath20 to estimate their contribution to the uncertainty in the anti - neutrino yield .
we also discuss the applications of our calculation to specific experimental configurations , including multiple reactor sources .
we draw our conclusions in section [ conclusion ] .
our analysis addresses issues common to large commercial power reactors , especially pressurized - water - reactors ( pwrs ) and boiling - water - reactors ( bwrs ) , using low - enrichment fuel .
explanation of the operation of pwrs and bwrs may be found in @xcite .
modern neutrino experiments are almost exclusively using @xmath1 from such reactors .
minor contributions from other reactor types are not considered here .
all uncertainties are given at the 68.3% confidence level .
the most accurate measurement of a reactor s thermal power is given by a calculation of the energy balance around the reactor vessel ( bwr ) or steam generator ( pwr ) .
this requires accurate measurements of feed - water mass flow and temperature , steam enthalpy and moisture content , and reactor coolant - cycle heat gains and losses @xcite . for the pwrs at diablo canyon power plant ( dcpp ) ( california , usa ) ,
the thermal balance is written as @xmath37 @xmath38 is the core thermal output , the time - integral of which gives @xmath28 .
the largest component of it , @xmath39 , is the power extracted from the steam produced directly in the steam generator .
smaller corrections are represented by : @xmath40 , the power lost in the water clean - up system ; @xmath41 , the power losses to the external environment due to radiation and convection ; @xmath42 , the contribution to @xmath43 from heating of the working fluid by the circulation pumps ; and @xmath44 , the contribution to @xmath43 from heating of the working fluid by the pressurizers .
the correction terms together account for 0.3 - 0.4% of the total @xmath38 .
a similar thermal balance may be written for other pwrs as well as for bwrs . in the case of bwrs ,
the main component is also in the form of steam produced by the reactor . for both reactor types
, @xmath43 is evaluated according to @xmath45 where @xmath46 is the mass flow rate of the feed - water to the steam generator ( pwr ) or reactor vessel ( bwr ) , and @xmath47 is the specific enthalpy rise in the steam generation .
we will discuss the uncertainties in these terms below .
the uncertainties on the terms other than @xmath39 in the thermal balance equation ( equation [ eq : eq_4 ] ) contribute negligibly to the uncertainty of the core thermal power and will not be addressed further here .
the enthalpy rise is calculated , using steam tables @xcite , from inlet and outlet values of the pressure and temperature , and from the moisture content on the secondary side of the steam generator or reactor vessel .
the enthalpy uncertainty has several contributions deriving from measured and calculated quantities .
the errors in the temperature and pressure measurements are assumed to be random after correction for known systematic contributions .
the moisture content of the steam is generally known to grow during the operation cycle because of reduced moisture removal efficiency of the steam separators due to a slow deposition of eroded metal particles .
the moisture content uncertainty of the steam may be treated as a systematic uncertainty since all reactors undergo this aging process .
the calculation of the enthalpy rise from these inputs using steam tables contributes an additional systematic uncertainty of @xmath480.2% @xcite . as an example
, table [ tab : table3 ] shows the uncertainties of the quantities relevant to the enthalpy calculation for dcpp .
the moisture content at dcpp s steam generators grows with time , as mentioned above ; the number in table [ tab : table3 ] is an average over the period of operation . over longer periods
the moisture content may grow , as explained above , by a factor of two over the level represented in table [ tab : table3 ] , but even then it will comprise a sub - dominant contribution to the error budget . a similar analysis for beaver valley unit 2 yielded enthalpy - related contributions to the thermal power uncertainty totaling 0.16% @xcite .
a more general discussion in @xcite uses an enthalpy uncertainty of 0.24% . for a generic reactor or when
a more detailed analysis is unavailable , we suggest conservatively assuming 0.15% random uncertainty ( for the @xmath49 and @xmath50 measurements ) and 0.2% systematic uncertainty ( due to the moisture content and use of steam tables ) for the enthalpy rise .
.[tab : table3 ] uncertainty contribution to the enthalpy from measured input quantities at dcpp .
the change of the enthalpy vs given quantity , ( @xmath51enthalpy/@xmath51quantity ) , is calculated using steam tables @xcite .
@xmath52 , @xmath53 and @xmath54 are the inlet pressure , the inlet temperature , and the outlet pressure , respectively .
the present moisture carryover value at dcpp is assumed in the table .
the quadratic sum conservatively assumes a full 0.2% error related to the use of steam tables to calculate the enthalpy rise . [ cols="<,^,^,^,^ " , ] the actual situation in the japanese reactor industry is probably somewhere in between the all - venturi case and the ideal case of all ufms .
many reactors may already be equipped with ufms , but some may not , and those that are may simply use them to calibrate their venturi meters . however , the total uncertainty , @xmath55 , does not change much between these two cases . thus for kamland ,
a reduced error of 0.6 - 0.8% may be more appropriate .
a better understanding of the nature of the japanese instrumentation and reactor simulations may result in yet smaller overall uncertainties for kamland .
the uncertainty in the estimated anti - neutrino interaction rate at reactor experiments has contributions from the thermal power and the fission calculation uncertainties .
the dominant contributions to the thermal power uncertainty come from the enthalpy rise and mass flow rate in the steam generator ( pwr ) or reactor vessel ( bwr ) .
the former has a systematic uncertainty of 0.2% and a random uncertainty of typically 0.15% .
the uncertainty in the mass flow rate depends on the instrumentation used to perform the measurement .
traditional venturi flow meters exhibit significant fluctuations of 1.4% and a systematic uncertainty of 0.6% due to fouling .
more precise ultrasonic flow meters have been evaluated to have systematic uncertainties below 0.5% and typically show random fluctuations below this level .
we estimated , for the first time to our knowledge , the contribution to the @xmath1 signal uncertainty from errors in the fission rate calculations .
these errors were extracted from comparisons of heavy element concentration measurements in spent fuel elements with calculations of those concentrations by the same simulations that reactor @xmath1 experiments rely on to obtain the fraction of fissions from each of four main heavy isotopes .
error propagation accounting for correlations between the four isotopes yielded a 0.1% systematic uncertainty on the @xmath1 signal rate , and a 0.9% random uncertainty .
we also estimated a 0.2% contribution due to the uncertainty in the measured energy release per fission of each isotope .
we demonstrated how these errors can be combined for typical reactor experiment configurations , and highlighted situations in which cancellations occur between different reactor sites . with this methodology
, we found that these contributions to the signal rate uncertainty in the kamland experiment may be reduced from 2.3% to 0.76% , or even 0.59% , pending better understanding of reactor instrumentation .
as we move deeper into the phase of precision reactor neutrino oscillation experiments , and in particular multi - reactor @xmath3 experiments , such sub - percent understanding of these errors and their systematic and random components will become more and more important . as reactor engineering , flow rate metrology , and core
modeling continue to improve , further reductions may be achievable in these uncertainties . toward this end , closer relationships between neutrino scientists and reactor operators should be encouraged .
we are grateful to harry miley for providing some of the references used here and to palo verde nuclear generating station for providing data used in the analysis .
we thank petr vogel and patrick decowski for their careful reading and useful comments .
we would also like to thank akira sebe and kazuhiro terao for their help in translating the japanese references .
this work was supported , in part , by u.s .
national science foundation grant no .
phy-0758118 , u.s .
department of energy contract no .
de - ac02 - 05ch1123 and grant nos .
de - fg02 - 01er41166 and de - fg02 - 04er41295 .
99 c. bemporad , g. gratta , and p. vogel , rev .
phys . * 74 * , 297 ( 2002 ) ; r. d. mckeown and p. vogel , phys . rept . * 394 * , 315 ( 2004 ) ; g. l. fogli _ et al._. , prog .
57 * , 742 ( 2006 ) ; m. c. gonzalez - garcia and m. maltoni , phys . rept . * 460 * , 1 ( 2008 ) . k. eguchi _
et al._. , phys .
rev . lett . * 90 * , 021802 ( 2003 ) .
k. eguchi _ et al .
* 94 * , 081801 ( 2005 ) .
s. abe _ et al .
lett . * 100 * , 221803 ( 2008 ) .
l. wolfenstein , phys .
d * 17 * , 2369 ( 1978 ) ; s. p. mikheyev and a. y. smirnov , sov .
. phys . * 42 * , 13 ( 1985 ) .
m. apollonio _ et al .
_ , eur . phys .
j. c * 27 * , 331 ( 2003 ) .
f. boehm _ et al .
d * 64 * , 112001 ( 2001 ) .
k. anderson _
et al . _ , hep - ex/0402041v1 ( 2004 ) .
f. ardellier _ et al .
_ , hep - ex/0606025v4 , ( 2006 ) .
x. guo _ et al .
hep - ex/0701029v1 ( 2007 ) .
j. c. anjos _ et al .
155 * , 231 ( 2006 ) .
s. b. kim _ et al . _ , http://neutrino.snu.ac.kr/reno ( accessed august 1 , 2008 ) .
s. barbeau , j. i. collar , and o. tench , jcap * 09 * , 009 ( 2007 ) .
a. bernstein _ et al .
_ , j. appl . phys .
* 91 * , 4672 ( 2002 ) ; a. bernstein _ et al .
_ , j. appl . phys .
* 103 * , 074905 ( 2008 ) . v.
i. kopeikin _ et al .
nuclei * 64 * , 849 ( 2001 ) .
l. miller , ph.d .
thesis , stanford university , 2001 .
k. schreckenbach _ et al .
b * 160 * , 325 ( 1985 ) .
a. a. hahn _ et al .
b * 218 * , 365 ( 1989 ) .
p. vogel , phys .
c * 76 * , 025504 ( 2007 ) .
p. vogel _ et al .
c * 24 * , 1543 ( 1981 ) .
y. declais _ et al .
b*338 * , 383 ( 1994 ) .
k. nakajima _ et al .
instrum . meth .
a * 569 * , 837 ( 2006 ) .
z. djurcic , ph.d .
thesis , university of alabama , 2004 .
j. k. shultis and r. e. faw , _ fundamentals of nuclear science and engineering _ ( crc press , boca raton , 2002 ) .
`` improved flow measurement accuracy using cross - flow ultrasonic flow measurement technology'' , westinghouse wcap-16163-p rev . 0 , 2004 .
w. t. parry _ et al .
_ , _ asme international steam tables for industrial use _ ( asme press , new york , 2000 ) . the international association for the properties of water and steam , `` revised release on the iapws industrial formulation 1997 for the thermodynamic properties of water and steam '' , 2007 , available at http://www.iapws.org/release.htm ( accessed august 1 , 2008 )
. `` calibration for beaver valley unit 2 lefm checkplus system '' , caldon inc .
ml129 , 2000 .
m. takamoto and a. andersson , `` what is the volume of one liter ? '' , iaea technical meeting on increasing power output and performance of nuclear power plants by improved instrumentation and control systems , prague , czech republic , 2007 .
v. r. foster jr .
flow control : the magazine of fluid handling systems , vol .
vii , no . 8 ( 2001 ) .
l. c. lynnworth , _ ultrasonic measurements for process control : theory , techniques , applications _
( academic press , san diego , 1989 ) . `` general electric bwr thermal analysis basis ( getab ) : data , correlation and design application '' , general electric co. nedo-10958-a , 1977 . `` caldon experience in nuclear feedwater flow measurements '' , caldon inc .
ml162 rev . 1 , 2002 . h. estrada , `` an assessment of the integrity and accuracy of feedwater flow and temperature measurements '' , caldon inc . tp11 rev . 2 , 1996 .
`` implications of power uprates on safety margins of nuclear power plants '' , iaea - tecdoc-1418 , 2004 .
us code of federal regulations , title 10 , part 50 , appendix k ( 2000 ) . ,
1986 ( in japanese ) . v. r.
foster jr . ,
flow control : the magazine of fluid handling systems , vol .
vii , no.6 , 16 ( 2001 ) .
`` improving thermal power accuracy and plant safety while increasing operating power level using the lefm system '' , caldon inc .
er-80p , rev . 0 , 1998 ( caldon inc
. proprietary ) .
h. estrada , `` identifying and bounding the uncertainties in lefm flow measurements '' , caldon inc.tp-13 , 1996 .
us nuclear regulatory commission , `` nrc staff position on use of the westinghouse crossflow ultrasonic flow meter for power uprate or power recovery '' , regulatory issues summaries 2007 - 24 ( 2007 ) .
d. w. spitzer , _ flow measurement _
( the instrumentation , systems , and automation society , research triangle park , 2001 ) .
`` nuclear feedwater flow measurement application guide '' , epri tr-112118 , 1999 .
n. furuichi , aist bulletin of metrology , vol .
ii , no . 3 , 569 ( 2005 ) ( in japanese ) .
n. furuichi _ et al .
_ , `` a facility with high reynolds number for calibration of a feedwater flowmeter '' , icone15 - 10209 , 2007 .
c. e. sanders and i. c. gauld , `` isotopic analysis of high - burnup pwr spent fuel samples from the takahama-3 reactor '' , ornl / tm-2001/259 , 2003 .
i.c.gauld , `` strategies for application of isotopic uncertainties in burnup credit '' , nureg / cr-6811 , ornl / tm-2001/257 , 2003
. `` user s manual for rocs '' , combustion engineering inc .
ce - ces-4-rev .
e. k. lee _ et al .
energy * 26 * , 983 ( 1999 ) .
k. suyama _ et al .
_ , j. nucl .
sci . tech .
* 39 * , 82 ( 2002 ) .
k. suyama , h. mochizuki , and t. kiyosumu , nucl . tech . * 138 * , 97 ( 2002 ) .
y. nakahara _ et al .
_ , nucl . tech . *
137 * , 111 ( 2002 ) .
r. j. guenther _ et al .
_ , pacific northwest national laboratory report no .
pnl-5109 - 104 , 1991 .
d. e. burk , m.s .
thesis , texas a&m university , 2005 .
`` standard test method for atom percent fission in uranium and plutonium fuel ( neodymium-148 method ) '' , american society for testing and materials , e321 , 1996 .
`` nuclear fuel data '' , energy information administration , form rw-859 , 2002 , available at http://www.eia.doe.gov/cneaf/nuclear/spent_fuel/ussnftab3.html ( accessed august 1 , 2008 ) .
o. w. hermann _ et al .
_ , `` validation of the scale system for pwr spent fuel isotopic composition analysis '' , ornl / tm-12667 , 1995 . o. w. hermann and m. d. dehart , `` validation of scale(sas2h ) isotopic predictions for bwr spent fuel '' , ornl / tm-13315 , 1998 . m. d. dehart and o. w. hermann , `` an extension of scale(sas2h ) isotopic predictions for pwr spent fuel '' , ornl / tm-13317 , 1996 .
`` benchmark of scale ( sash2 ) isotopic predictions of depletion analysis for san onofre pwr mox fuel '' , ornl / tm-1999/326 , 1999 .
y. nakahara , k. suyama , and t. suzuki , translation of `` technical development on burn - up credit for spent lwr fuel '' , jaeri - tech 2000 - 071 , 2000 [ ornl / tr-2001/01 , 2001 ] .
m. d. dehart , m. c. brady , and c. v. parks , `` oecd / nea burnup credit criticality benchmark phase i - b results '' , ornl-6901 , 1996 . y. nakahara _ et al .
_ , radiochim .
acta * 50 * , 141 ( 1990 ) .
g. zacek _ et al .
d * 34 * , 2621 ( 1986 ) .
y. declais _ et al .
b * 338 * , 383 ( 1994 ) . | anti - neutrino emission rates from nuclear reactors are determined from thermal power measurements and fission rate calculations .
the uncertainties in these quantities for commercial power plants and their impact on the calculated interaction rates in @xmath0 detectors is examined .
we discuss reactor - to - reactor correlations between the leading uncertainties , and their relevance to reactor @xmath1 experiments . |
one of the most important aims of particle physics is the test of models , particularly the standard model . testing the standard model means , on one hand , to increase the accuracy of its parameters and , on the other hand , the search for deviations of its predictions from measured quantities to discover new effects . to do this experiments
are accomplished with increasing accuracy of measurement .
therefore theoretical calculations have to be performed with increasing numerical precision .
a parameter of the standard model of particular interest is the strong coupling constant @xmath1 .
it can , for example , be determined via the so - called @xmath2-ratio , the normalized total cross section of the process @xmath3 , @xmath4 in order to achieve high precision one has , within perturbative quantum field theory , to calculate high orders of the perturbation series .
this means , we need to calculate large numbers of more and more complicated feynman diagrams . therefore different methods for the evaluation of such multiloop feynman diagrams have been developed .
( see e.g. @xcite . ) the quantity @xmath5 is nowadays known to @xmath6 precision @xcite and first results in @xmath7 have been published @xcite . for
a review see refs .
@xcite .
a prescription to solve problems with a large number of complicated feynman diagrams is to reduce them to a set of master integrals , e.g. by means of the traditional integration by parts method , and then evaluate these .
the aim of the project described in this article is to develop a method for the calculation of massless integrals .
more precisely , we want to consider the planar ( @xmath8 ) and non - planar ( @xmath9 ) three - loop propagator diagrams : @xmath10 @xmath11 they can be considered as the master integrals of the corresponding class of feynman diagrams , which are `` building blocks '' of the calculations for @xmath5 in @xmath7 . + we are interested in the @xmath12-expansion of the diagrams , where @xmath12 is the parameter of _ dimensional regularization_. this means , the integrals are formally calculated in @xmath13 dimensions and then expanded in @xmath12 . by means of this procedure
uv divergencies of diagrams manifest themselves as poles in @xmath12 and can then be cancelled by means of _
renormalization_. it is thus essential for renormalization to know the pole structure of a feynman diagram and the finite part of the expansion .
however , when one has to deal with products of graphs with poles as in the case of the following diagram , which appears within the mentioned calculations of @xmath5 , one has to calculate higher @xmath12 orders of the second factor in order to ensure , that the result is correct up to the finite term .
this is why we are interested in higher orders of the @xmath12 expansion of the diagrams under consideration .
the @xmath14 terms of the diagrams @xmath9 and @xmath8 are necessary for the computation of the @xmath15 function of the scalar @xmath16 model at five - loop level . they have been known since long from ref .
@xcite and @xcite .
the calculation of the planar diagram is somewhat simpler than that of the nonplanar one , because it could be related to the generic massless two - loop two - point function by means of integration by parts @xcite .
the latter was studied in many works .
e.g. @xcite and references therein .
recently , in ref . @xcite it was in a sense computed analytically .
thus one can say that the planar diagram is known to arbitrary order and certainly to sufficient precision for practical needs . the known results for the diagrams @xmath8 and @xmath9 read @xcite @xmath17 @xmath18 and @xmath19 @xmath20 in the following we denote the expansion coefficients of order @xmath21 by @xmath22 , e.g. @xmath23 .
these numbers have been used to check the method and to examine the properties of the numerical algorithms described later .
the determination of the @xmath24 coefficient of @xmath9 is one of the aims of the project presented in this paper .
an independent evaluation can be found in ref .
@xcite , the result was given in ref .
@xcite . a numerical approach to the calculation of integrals of this type
is given e.g. in ref .
@xcite .
the paper is organized as follows : section [ gpxt ] shows how a representation of the diagrams @xmath9 and @xmath8 in terms of sums can be found by means of the gegenbauer polynomial method .
section [ sums ] gives an overview over the harmonic sums and algorithms to deal with them .
it is shown how the summation algorithms are applied to evaluate the sums representing the diagrams . for the @xmath9 diagram an analytical solution by means of the described algorithms
is not yet possible . in this case
one can get numerical results by summing up the series .
section [ convergence ] reviews methods to accelerate the convergence of such processes to get better results . by means of these numerical methods it is possible to calculate the expansion coefficients with high precision . for the non - planar diagram
this is described in section [ nonplanar ] .
the planar topology can be calculated analytically with the help of the summation algorithms , which is demonstrated in section [ planar ] .
the gegenbauer polynomial x - space technique ( gpxt ) to evaluate feynman diagrams in coordinate space was introduced in ref .
applications can be found e.g. in refs .
@xcite , an extension of the technique in ref . @xcite . in this section
we will apply the method to the diagram @xmath9 .
the results for the planar diagram can be obtained analogously .
labeling the momenta as shown in fig . [ n_labels ] the integral representation of @xmath9 reads [ n_labels],height=188 ] @xmath25 by means of fourier transformation @xmath26 where @xmath27 , the integrand is transformed to position space .
the integrals over the momenta can then be evaluated to delta functions : @xmath28 after the evaluation of the delta functions and choosing an appropriate coordinate system ( see fig .
[ n_labels ] , where the root vertex is denoted by 0 ) the denominators get a simple and uniform shape : @xmath29 spherical coordinates are introduced by means of @xmath30 where @xmath31 now the propagator terms and the exponential function can be expanded in gegenbauer polynomials using formulae ( [ ggp - entwicklung ] ) and ( [ exp - entwicklung ] ) of appendix [ anhang_gp ] .
this leads to a representation of @xmath9 as a multiple sum @xmath32 where @xmath33 the angular integrations can be evaluated using the properties of the gegenbauer polynomials .
each application of the orthogonality relation ( [ ggp_ors ] ) reduces two @xmath34 to one and produces a kronecker delta .
finally , one arrives at the expression @xmath35 the complicated structure of the triple sum is introduced by eq .
( [ cc - summe ] ) , which is applied when there are two polynomials with the same arguments . to handle the radial integrals the region of integration
is distributed in 120 regions according to the relative size of @xmath36 . in each region
the expressions @xmath37 , eq .
( [ m_def ] ) , evaluate to rational expressions . for simplification @xmath38
is set to 1 ; the @xmath39-dependence can later be restored by dimensional arguments .
thus we are lead to integrals like @xmath40 @xmath41 actually , due to the symmetry of the diagram , only 30 of these 120 terms are different .
the calculation of these can be automated . after performing all integrations and rewriting @xmath42
the integral @xmath9 is expressed in terms of a triple sum : @xmath43 @xmath44 where @xmath45 @xmath46 finally , according to ref .
@xcite , this result can be cast into the so - called @xmath47-form . rewriting @xmath48
one obtains @xmath17 @xmath49 @xmath50 can now be expanded in powers of @xmath12 . for each order in @xmath12 a triple sum
has to be evaluated .
the calculation of these sums is the aim of the rest of the paper .
in this section the harmonic sums are introduced and their relationship to the problem under consideration is shown .
harmonic sums are defined as @xcite @xmath51 which , for @xmath52 , leads to riemann s zeta function : @xmath53 higher harmonic sums are defined @xcite recursively via the relation @xmath54 the number of indices of a harmonic sum is called its _ depth _ , the sum of the absolute values of the indices its _ weight_. a negative index denotes an alternating sum , by overlining the index . in this case
a negative sign of the index indicates consequently a positive power of the corresponding variable . ]
@xmath55 in ref .
@xcite different algorithms are described to deal with expressions of the form @xmath56 and sums over such .
these are implemented in the program package summerwritten in form @xcite which is available under the url + http://www.nikhef.nl/~form/formdistribution/packages/summer .
the harmonic sums are related to the psi ( digamma ) function , which is the logarithmic derivative of the gamma function , and its derivatives , the polygamma functions @xmath57 , @xmath58 @xmath59 via the following relations @xcite : @xmath60 @xmath61 . \label{psin->s}\ ] ] using these relations the gamma function can be expanded in terms of harmonic sums ( cf .
@xcite ) @xmath62 the expansion of the exponential function leads to the expansion formulae used in the following : @xmath63 \nonumber \\ & & + { \textstyle \frac{1}{6}}\ , { \varepsilon}^3 \left [ s_1 ^ 3(n-1 ) - 3 s_1(n-1 ) s_2(n-1 ) + 2 s_3(n-1 ) \right ] \nonumber \\ & & + { \mathcal{o}}({\varepsilon}^4 ) , \nonumber \\ \label{1/gamma - s - entwicklung } \frac{\gamma(1+{\varepsilon})\,\gamma(n)}{\gamma(n+{\varepsilon } ) } & = & 1 - { \varepsilon}s_1(n-1 ) \\ & & + { \textstyle \frac{1}{2}}\ , { \varepsilon}^2 \left [ s_1 ^ 2(n-1 ) + s_2(n-1 ) \right ] \nonumber \\ & & - { \textstyle \frac{1}{6}}\ , { \varepsilon}^3 \left [ s_1 ^ 3(n-1 ) + 3 s_1(n-1 ) s_2(n-1 ) + 2 s_3(n-1 ) \right ] \nonumber \\ & & + { \mathcal{o}}({\varepsilon}^4 ) .
\nonumber\end{aligned}\ ] ] the structure of the sum given in eq .
( [ ns ] ) , @xmath64 particularly the boundaries of the innermost sum , is very difficult to handle with symbolic manipulation programs .
therefore the first step is the simplification of this sum . for a given @xmath65 we have to sum over pairs of indices @xmath66 , which satisfy @xmath67 neglecting the condition @xmath68 at first
, this is equivalent to @xmath69 which is described by the rectangular area in the @xmath66-plane shown in fig .
[ summationsdiagramm ] . .
[ summationsdiagramm ] ] taking into account the condition @xmath68 this means , that the sum ranges over all combinations of indices in this marked area whose sum is even .
let us in a next step consider pairs @xmath66 with constant sum @xmath70 , i.e. points on straight lines parallel to the second bisector .
the sum starts with the line @xmath71 ( @xmath72 ) and reaches over the shaded rectangle in fig .
[ summationsdiagramm ] to infinity . taking into account the symmetry of the terms
it is sufficient to restrict oneself to the area below or over the first bisector .
the contributing points on such a line can be described by the condition @xmath73 , where @xmath74 ( cf .
[ summationsdiagramm ] ) .
if we parametrize @xmath75 and @xmath76 , the addends read @xmath77 and @xmath78 , respectively , and it is obvious , that in the second case the sum of the arguments can never be even , while the terms of the first kind automatically fulfill the condition .
therefore the original sum is transformed to a much simpler form : @xmath79 if the expansions ( [ gamma - s - entwicklung],[1/gamma - s - entwicklung ] ) are applied to the sum in eq .
( [ 7fachsumme ] ) , it is represented as a triple sum over rational expressions and products of such and harmonic sums . in the following some algorithms
are discussed to simplify such objects . to sum expressions of the form @xmath80 where @xmath81 is a polynomial in @xmath82 , one first performs partional fractioning .
this leads to simpler summands in which @xmath82 appears only once in the denominator .
the resulting sums can be evaluated easily in terms of zeta or psi functions by means of the formula @xcite @xmath83 if the upper limit is infinity , the second psi function is zero for @xmath84 . in ref .
@xcite different algorithms have been published to evaluate nested sums using the properties of higher harmonic sums and more general structures like the so - called @xmath85- and @xmath86-sums . for our purpose
some special cases of the algorithms a and b of this work are needed .
if an addend is not of the simple form @xmath87 , but contains an offset in the denominator , which must not be symbolic , the summand can be reduced recursively @xmath88 this is a special case of algorithm a. of course , this is also valid for sums of depth greater than 2 .
algorithm b describes the recursive simplification of sums where products of harmonic sums appear .
an expression of the form @xmath89 can be rewritten by inserting the definition of @xmath90 .
then shifting variables , reordering of summations and partial fractioning yields to expressions which can be evaluated immediately and such of the same form as the original expression to which the procedure is applied again .
thus the sum is reduced recursively to simpler expressions and finally evaluated . in ref .
@xcite algebraic relations between harmonic sums have been studied . for harmonic sums with identical arguments
the permutation relation holds : @xmath91 where @xmath92 this allows us to express the product of two harmonic sums in the following way : @xmath93 for products with higher harmonic sums one has relations like @xmath94 more relations of this kind can be found in @xcite . in the following
also sums with double argument @xmath95 will appear .
these can not be handled with the algorithms described up to now . in ref .
@xcite transformations of such objects to simpler sums have been discussed . from this work
we need the relation @xmath96 which relates the sum over @xmath97 to two sums over @xmath98 , one of which is alternating .
the opposite operation , the transformation of a sum with simple argument to one with double argument , can be done by means of a special case of the refinement formula @xcite @xmath99 generally , we can write a sum of the form @xmath100 in the following way : @xmath101 . \label{sumf(2x)}\ ] ] the sum over all even numbers is replaced by the sum over all integers minus the sum over the odd numbers .
this is achieved by subtracting an alternating sum and the division by 2 . finally , we need some formulae to convert nested sums to multiple infinite sums or vice versa .
two important relations are given by @xmath102 @xmath103 eq .
( [ sum2l ] ) can be obtained by splitting the sum on the left - hand side into two sums , which is visualized in fig .
[ sum2l_1 ] , and paramaterizing the summation variable @xmath104 and @xmath105 , respectively .
this leads to sums of the type @xmath106 .
these can be converted to a double infinite series by shifting the inner summation variable . ) in two sums [ sum2l_1 ] ] if the argument of the appearing harmonic sums is a sum of two or more variables , the following relations are helpful : @xmath107 @xmath108
if the sums can not be completely simplified analytically , one can at least get a numerical result by summing up the terms with the computer .
however , in our case the sum converges very slowly . in this section
we briefly sketch the method of _ nonlinear sequence transformations _ , by means of which one can improve the convergence of a series . for further information
we refer to ref . @xcite and references cited therein .
we consider a sequence @xmath109 and the corresponding infinite series @xmath110 with the partial sums @xmath111 which themselves form a sequence @xmath112 that converges to the value s. the question is now , whether a different sequence @xmath113 ( or @xmath114 ) exists that converges to the same value @xmath85 but has a better convergence behaviour . the sequence @xmath113 is said to _ converge faster _ than @xmath112 to the common limit @xmath85 , if @xmath115 in practice this means that @xmath116 is a better approximation to the exact value of the limit than @xmath117 , if the partial sum of order @xmath9 is known .
convergence acceleration methods can be found with the help of model sequences .
this means , one investigates a particular class of sequences and constructs a transformation , i.e. a prescription how to calculate the elements of a new sequence @xmath113 out of the elements of @xmath112 , which is then exactly valid for the considered sequences .
the prescriptions one has found in this way are often able to improve the convergence of different sequences of similar structure .
one of the oldest and best known methods is _
aitken s @xmath118-process_. the partial sums are transformed according to the rule @xmath119
@xmath120 is the _ forward difference operator _ , which is defined as @xcite @xmath121 powers of @xmath120 are to be understood as multiple application of @xmath120 , e.g. @xmath122 = \delta \left [ f(n+1 ) - f(n ) \right ] \nonumber \\ & = & f(n+2 ) - 2 f(n+1 ) + f(n).\end{aligned}\ ] ] another method which can be found in the literature on numerical methods is _
wynn s epsilon algorithm _ , which is defined by the recursion scheme @xmath123 we follow the notation of ref .
the lower index indicates the transformation order , the upper index labels the sequence elements .
the difference operator acts only on the upper index .
similar to the epsilon algorithm is _
wynn s rho algorithm_:@xmath124 where the _ interpolation points _
@xmath125 have to fulfill @xmath126 for @xmath127 and the sequence @xmath128 diverges to infinity .
a possible choice is @xmath129 .
an example , where the rho algorithm is very powerful , is the series defining @xmath130 .
when one sums the series up to the order @xmath131 one gets only 4 correct decimal digits .
application of the rho algorithm delivers more than 250 correct digits . for the _ levin transformations _
remainder estimations have to be taken into account .
the levin transformation ist exact for sequences of the form @xmath132 where @xmath85 is the limit of the sequence and @xmath133 is a parameter which can , in principle , be chosen arbitrarily and is usually set equal to 1 .
@xmath134 are the remainder estimates .
they are , of course , generally not known exactly , because the actual limit is unknown .
therefore one uses assumptions for the @xmath134 which are customized to the given problem and which define different kinds of levin transformations . as an example , the ansatz @xmath135 yields _ levin s v transformation _ @xmath136 an enhancement of the epsilon algorithm is _
@xmath137 algorithm_:@xmath138 \left [ \delta \vartheta_{2k+1}^{(n+1 ) } \right]}{\delta^2 \vartheta_{2k+1}^{(n ) } } , \qquad k , n \in { \mathds{n}_0}. \nonumber\end{aligned}\ ] ] it is now suggestive to apply such a procedure , that improves the convergence of a series , succesively . in ref .
@xcite algorithms are discussed which allow the calculation of high transformation orders efficiently and with low memory costs .
the principle is to calculate for each new element of the series all possible elements of the matrix @xmath139 .
so one always gets the highest possible transformation order .
details can be found in ref .
@xcite , where also many more acceleration methods are discussed . for this work
the algorithms mentioned above were programmed in fortran . to achieve the desired numerical precision we used bailey s library mpfun @xcite , which allows calculations with arbitrary precision .
in this section we sketch the application of the described methods to the nonplanar diagram @xmath9 and show how a numerical result is obtained .
the sum we have to calculate reads ( [ ns ] ) @xmath140 @xmath141 with the addend terms @xmath142 where @xmath143 consists of rational expressions resulting from the radial integration of eq .
( [ radialintegral ] ) .
the coefficient of @xmath144 is a triple sum of rational expressions , where no harmonic sums occur .
the sum can be evaluated analytically , which we sketch briefly in this subsection .
the sum over @xmath145 can be performed easily and we are left with a double sum . in the summands
we get harmonic sums with arguments @xmath146 and @xmath147 .
the resulting rational terms and terms with @xmath148 can now be summed over @xmath21 .
the terms containing @xmath149 can be cast into an expression that can be summed over @xmath146 by means of a transformation of variables . after the summation
we are left with a simple sum with increased number of terms .
the summands now consist of rational expressions , harmonic sums with simple argument @xmath150 , such with double argument @xmath151 and products of the latter .
again , the rational expressions can be summed up analytically by means of mathematica or summer .
expressions containing harmonic sums with simple arguments can be summed with summer .
expressions with harmonic sums with double argument can be transformed to terms calculable with summer by means of formula ( [ s2x ] ) .
a bit more complicated are terms where @xmath152 appears in the denominator , like @xmath153 by means of eq .
( [ s(n)->s(2n ) ] ) the harmonic sum is converted to sums with double argument @xmath154.\ ] ] then the sum is rewritten according to eq .
( [ sumf(2x ) ] ) @xmath155\ ] ] and can now be calculated . finally , there are the products @xmath156 which can be evaluated with the help of eq .
( [ s(n)->s(2n ) ] ) .
the problem for the @xmath144 coefficient can thus be solved completely and the result reads @xmath157 which agrees with @xcite . in the expression for the @xmath24 coefficient the terms in the triple sum contain the harmonic sums @xmath158 , @xmath159 , @xmath160 and @xmath161 .
all expressions can be summed over a variable which does not appear as an argument in @xmath85 .
the result is then a double sum @xmath162 where terms with harmonic sums of the following types occur : @xmath163 in the case of @xmath164 and @xmath165 we sum now over the variable @xmath21 which is not the argument of @xmath85 .
expressions containing @xmath166 , @xmath167 can be handled by means of variable transformations as described in section [ nestedsums ] .
the terms with @xmath168 , @xmath169 can in the most cases be cast into a form to which algorithm b of ref .
@xcite can be applied .
thus the second summation yields a simple sum .
now the addends contain simple harmonic sums and products of two or three harmonic sums , some of which have double arguments : @xmath170 these terms can in principle be calculated by means of the described algorithms
. however , in some of the terms there are certain combinations of denominators and harmonic sums in which the summation variable is doubled .
these can in some cases be be handled with the known algorithms , however , not in the general case .
therefore the problem of the order @xmath24 can not be solved analytically at present . to achieve this , more examinations in the subject of the summations are necessary . since up to
now an analytical solution is not possible with our method we try to find a numerical solution . to do this
, we simplify the terms as far as possible and calculate the sums numerically . for the numerical calculation the terms
have been simplified with mathematica and output in fortran syntax .
this code has been optimized by means of maple and built in a fortran program . to speed up the calculation
this program was parallelized by means of mpi @xcite .
finally , the program was translated with bailey s program transmp @xcite to allow arbitrary numerical precision by means of the mpfun library @xcite .
the programs store all partial sums . to the sequence of partial sums
we apply different convergence acceleration methods .
a different approach to a numerical solution , is the following : we begin with the representation of the sum ( [ formsumme ] ) @xmath171 for a fixed @xmath65 the expression @xmath172 describes @xmath173 infinite sums @xmath174 sums of that kind can easily be calculated with summer .
so the obvious procedure is to evaluate all contributions @xmath175 for all values of @xmath21 up to an upper limit @xmath176 .
thus we get a sequence of expressions consisting of rational numbers and zeta functions .
these can be evaluated numerically to achieve a sequence of decimal numbers @xmath112 , where @xmath177 to this we again apply the convergence acceleration algorithms .
the first elements of this sequence are @xmath178 and the 500th element is ( with rounded coefficients ) @xmath179 this procedure can be implemented easily in form or its parallelized version parform @xcite .
we show in table [ tab_e2_2 ] the results of the fortran - program .
the entry in the line labeled ` sum ' is the result we get by simply summing up the terms , where the upper limit of the outermost summation was set to 15,000 .
the other lines contain the results obtained by application of different algorithms .
.numerical results for the @xmath24 coefficient of @xmath9 [ tab_e2_2 ] [ cols="<,>",options="header " , ] one can see , that the results of the aitken and the rho method agree in 21 decimal digits .
these algorithms delivered the best result .
application of the rho algorithm yields 13 correct digits in this case . ] for the numerical calculations of the planar topology , where the exact result was known and could be used to check the results . since the structure of the terms is similar , one can expect , that they are also suitable for the nonplanar diagram .
the result of the epsilon procedure agrees in 20 digits with the former .
all results agree in 16 digits . in a very pessimistic interpretation
at least these digits can be considered reliable .
the results of the numerical evaluation by means of the form - method agree with these numbers in 14 or 20 digits , depending on the acceleration method used . finally , we want to get an idea of the analytical result by means of `` experimental mathematics '' .
looking at the known result for the planar diagram , we assume , that the result for @xmath9 be also a linear combination of zeta values with integer coefficients , which have to be determined
. this problem is known as _ integer relation detection _
@xcite .
in general a vector @xmath180 with real or complex elements is said to possess an _
integer relation _ , if integers @xmath181 exist such that at least one of them is different from zero and @xmath182 an algorithm which can solve this problem is the _ pslq algorithm _ @xcite .
the input for this algorithm are in our case the numerically calculated number and numerical values for the constants , i.e. the zeta values .
then the program checks , whether there is an integer relation between the numbers within the desired accuracy .
we used an implementation of this method by o. veretin @xcite .
assuming that the transcendentals are the same as for the planar diagram , as it is the case in the lower orders in @xmath12 , we find the following result for the 2nd order coefficient of the @xmath12 expansion of @xmath9 : @xmath183 this agrees with the analytical result which was found in the meantime @xcite . comparison of this result with table [ tab_e2_2 ] shows , that all the digits printed for the rho algorithm are correct .
this proves the power of this algorithm for our problem and justifies the confidence in the convergence acceleration techniques retrospectively .
in this section we show that the coefficient of @xmath24 for the planar diagram can be found analytically by means of the described method .
the steps we take are the same as described in section [ nonplanar ] .
the structure of the sums and of the terms are similar and so is the procedure of the calculation .
the difference is , the terms that cause problems in the case of @xmath9 , because we do not know algorithms for their evaluation , do not appear in the case of @xmath8 .
after application of the algorithms we have the sum in the form @xmath184 where @xmath185 is the contribution for @xmath186 , @xmath187 and @xmath188 @xmath189 @xmath190 consists only of terms that can be calculated with summer .
doing this we arrive at the known result @xmath191
we described a method for the calculation of massless , dimensionally regularized feynman integrals in their @xmath12-expansion .
it is based on the expansion of the integrand in terms of gegenbauer polynomials .
this yields sums over gamma functions which can be expanded in harmonic sums . as a consequence one arrives at sums over rational functions and products of such with harmonic sums which can be simplified by the introduced summation algorithms and then be calculated numerically or analytically .
the procedure in principle is independent of the gegenbauer polynomial technique .
it can be useful wherever there are sums of the described form .
the method should be expandable to more complicated problems , e.g. higher orders of the diagrams under consideration or more complex diagrams . in the case of the nonplanar diagram we could not get to a complete analytical solution . to achieve this , further knowledge in the summation techniques is required .
for this work we combined the abilities of different computer programs , in particular the pattern matching facilities of mathematica , the code generation functions of maple , and form s power of managing very large expressions .
it is the idea to implement the method in one programming language to make it stand - alone .
this could either be done in form or in c++ .
the summation algorithms of ref .
@xcite have already been implemented in both , the form version has recently been published @xcite . in @xcite also a mathematica implementation was introduced .
the coefficients of the @xmath12 expansion of the nonplanar three - loop - diagram @xmath9 could be calculated numerically to 20 digits @xmath192 the analytical result for the @xmath193 term given in @xcite could be reproduced , @xmath194 application of the pslq algorithm to the @xmath195 term gives @xmath196 which agrees with @xcite . also the analytical result up to order @xmath24 for the planar diagram was verified .
the author wishes to thank k. chetyrkin for the suggestion to do this work , for his support and information about gpxt and multiloop calculations .
i thank p. uwer , o. veretin and s. weinzierl for useful discussion about the harmonic sums and related topics .
i thank o. veretin for providing me with his implementation of the pslq algorithm .
e. j. weniger from the university of regensburg provided me with extensive information about convergence acceleration methods .
i am grateful for inspiring correspondence in this subject .
finally the author wishes to thank m. steinhauser and k. chetyrkin for reading the manuscript and many useful comments .
this work was supported by the graduiertenkolleg `` hochenergiephysik und teilchenastrophysik '' and by the sonderforschungsbereich tr 9 , `` computational particle physics '' .
the _ gegenbauer polynomials _ , also called _ ultraspherical polynomials _ , are the coefficients in the power series exansion of the generating function @xmath197 m. tentyukov , d. fliegner , m. frank , a. onischenko , a. rtey , h. m. staudenmaier , j. a. m. vermaseren : _ parform : parallel version of the symbolic manipulation program form _ * in : * v. g. ganzha , e. w. mayr , e. v. vorozhtsov ( eds . ) : _ computer algebra in scientific computing _ ,
proceedings of the seventh international workshop on computer algebra in scientific computing , july 1219 , 2004 , st .
petersburg , russia ; cs.sc/0407066 . | a method for the evaluation of the @xmath0-expansion of multi - loop massless feynman integrals is introduced .
this method is based on the gegenbauer polynomial technique and the expansion of the gamma function in terms of harmonic sums .
algorithms for the evaluation of nested and harmonic sums are used to reduce the expressions to get analytical or numerical results for the expansion coefficients . methods to increase the precision of numerical results are discussed .
ttp05 - 08 + sfb / cpp-05 - 16 multi - loop feynman integrals , gegenbauer polynomial technique , harmonic sums , nested sums , convergence acceleration 02.70.wz , 12.38.bx |
the cp asymmetries in the hadronic decays of b and k mesons involve strong final state phases .
thus strong interactions in these decays play a crucial role .
the short distance strong interactions effects at quark level are taken care of by perturbative qcd in terms of wilson coefficients .
the ckm matrix , which connects the weak eigenstates with mass eigenstates , is another aspect of strong interactions at quark level . in the case of semi
leptonic decays , the long distance strong interaction effects manifest themselves in the form factors of final states after hadronization .
likewise the strong interaction final state phases are long distance effects .
these phase shifts essentially arise in terms of s - matrix which changes an in state into an out state viz .
@xmath25 in fact , the cpt invariance of weak interaction lagrangian gives for the weak decay @xmath26
@xmath27 taking out the weak phase @xmath28 , the amplitude @xmath6 can be written as @xmath29 then eq .
@xmath30 implies @xmath31 it is difficult to reliably estimate the final state strong phase shifts .
it involves the hadronic dynamics .
using isospin , c - invariance of s - matrix and unitarity , we can relate these phases . in this regard ,
following cases are of interest : case ( i ) : the decays @xmath32 described by two independent single amplitudes @xmath6 and @xmath33 with different weak phases : @xmath34 where the states @xmath35 and @xmath36 are c conjugate of each other such as states @xmath37 , @xmath38 , @xmath39 . for case ( i )
, there is an added advantage that the decay amplitudes @xmath6 and @xmath21 are given by tree graphs .
assuming factorization for tree amplitudes , it is shown that the form factors @xmath40 , @xmath41 , @xmath42 obtained from the experimental branching ratios are in excellent agreement with heavy quark effective theory ( hqet ) .
hence factorization assumption is experimentally on sound footing for these decays .
case ( ii ) : the weak amplitudes @xmath43 , @xmath44 \\ a_{\bar{f } } & = \langle \bar{f}\left| \mathcal{l_{w}}\right| b^{0}\rangle = % \left[e^{i\phi_{1}}f_{1\bar{f } } + e^{i\phi_{2}}f_{2\bar{f } } \right]\end{aligned}\ ] ] as is the case for the following decays , @xmath45 the @xmath46 invariance of s - matrix gives @xmath47 which implies @xmath48
the time reversal invariance gives @xmath49 where @xmath50 is the weak interaction lagrangian without the ckm factor such as @xmath51 . from eq .
@xmath52 , we have @xmath53 it is understood that the unitarity equation which follows from time reversal invariance holds for each amplitude with the same weak phase . above equation can be written in two equivalent forms : 1 . exclusive version of unitarity @xcitewriting @xmath54we get from eq ( [ 04 ] ) , @xmath55 + where @xmath56 is the scattering amplitude for @xmath57 and @xmath58 is the decay amplitude for @xmath59 . in this version ,
the sum is over all allowed exclusive channels .
this version is more suitable in a situation where a single exclusive channel is dominant one . to get the final result ,
one uses the dispersion relation . in dispersion relation two particle unitarity gives dominant contribution . from eq.([06 ] ) , using two particle unitarity , we get @xcite , @xmath60where @xmath61 , @xmath62 eq.@xmath63 is especially suitable to calculate rescattering corrections to color suppressed @xmath64-amplitude in terms of color favored @xmath64-amplitude as for example rescattering correction to color suppressed decay @xmath65 in terms of dominant decay mode @xmath66 . before using two particle unitarity in this form , it is essential to consider two particle scattering processes .
+ @xmath67 or @xmath0 and @xmath1-invariance of @xmath68-matrix can be used to express scattering amplitudes in terms of two amplitudes @xmath69 and @xmath70 which in terms of regge trajectories are given by @xcite@xmath71for linear regge trajectories , using exchange degeneracy , we have @xmath72we take @xmath73 , @xmath74gev@xmath75 @xmath76gev@xmath77 . using @xmath78 and taking @xmath79
we get @xmath80 @xmath81 hence for @xmath82 or @xmath83 scattering we get @xmath84where @xmath85 + for @xmath86 , @xmath87@xmath88from eq.([6a ] ) and ( [ 6f ] ) with the use of dispersion relation , we obtain @xmath89we get @xmath90 by putting @xmath91 in @xmath92 .
now @xmath93 hence with rescattering correction @xcite @xmath94 \label{6h}\end{aligned}\]]where @xmath95 hence the final state phase shift @xmath96 for the color suppressed amplitude induced by the final state interaction is given by @xmath97with @xmath98 which we get from @xmath99for @xmath100 the color suppressed @xmath64-amplitude with rescattering correction is given by @xmath101 \label{6k}\]]where @xmath102 @xcite .
hence @xmath96 generated by the final state interaction is given by @xmath103 + to conclude : the scattering amplitude @xmath104 for the two particle final state obtained in eq.@xmath105 is used in the unitarity equation to generate the final state strong phase by rescattering for the color suppressed tree amplitude .
2 . inclusive version of unitarity [ 2]this version is more suitable for our analysis . for this case , we write eq . in the form@xmath106 parametrizing s - matrix as @xmath107@xcite , @xmath108 we get after taking the absolute square of both sides of eq.([07a])@xmath109 = \sum_{n , n^{\prime } } f_{n}s_{nf}^{\ast } f_{n^{\prime } } ^{\ast } s_{n^{\prime } f } \label{08}\ ] ] the above equation is an exact equation . in the random phase approximation
[ 2 ] , we can put @xmath110 we note that in a single channel description @xcite : @xmath111 the absorption takes care of all the inelastic channels .
similarly for the amplitude @xmath112 , we have @xmath113 the c - invariance of s - matrix gives : @xmath114 thus in particular c - invariance of s - matrix gives @xmath115 hence from eq .
@xmath116 , using eqs .
@xmath117 , we get @xmath118=\rho ^{2},\bar{\rho}^{2 } \label{13}\ ] ] where @xmath119from eq.([13 ] ) , we get @xmath120 ^{1/2 } \label{15}\]]the maximum value for @xmath121 is 1 and the minimum value for them is @xmath122 hence we get the following bounds:@xmath123from now on , we will confine our self to positve square root in eq,([15 ] ) .
the strong interaction parameter @xmath124 and @xmath125 in the above bounds can be obtained from the scattering amplitude @xmath126 given in eq.(12 ) obtain from regge pole analysis .
the @xmath127wave scattering amplitude @xmath128 is given by @xmath129for the scattering amplitude @xmath130 relevant for @xmath131 and @xmath132 , we obtain from eq.([30 ] ) using eq.(12 ) @xmath133
\label{31b}\end{aligned}\]]where we have used @xmath134 gev@xmath135 . for @xmath136
we have used the values of reference [ 2 ] whereas for @xmath137 and @xmath138 for @xmath139 and @xmath140 respectively . using the relation @xmath141 where @xmath128 is given by @xmath142 , the phase shift @xmath143 the parameter @xmath125 and the phase angle @xmath144 can be determined .
one gets @xmath145@xmath146hence we get the following bounds@xmath147 further we note that for these decays , @xmath148-quark is converted into @xmath149 or@xmath150 quark : @xmath151 . in particular for the tree graph ,
the configuration is such that @xmath152 and @xmath153 essentially go together into a color singlet state with the third quark @xmath154 recoiling ; there is a significant probability that the system will hadronize as a two body final state @xcite .
this physical picture has been put on the strong theoretical basis @xcite , where in these references the qcd factorization have been proved . for the tree amplitude , factorization implies @xmath155 we , therefore take the point of view that effective final state phase shift is given by @xmath156 we take the lower bound for the tree amplitude so that final state effective phase shift @xmath155 thus for @xmath157 the decay @xmath158 is described by two amplitudes @xcite@xmath159 = \left\vert p\right\vert \left [ 1-re^{i(\gamma + \delta _ { + -})}\right ] \label{16e}\]]where @xmath160the decay @xmath161 is described by the two amplitudes @xcite @xmath162 \label{16f}\]]where @xmath163for these decays , we use the lower bounds in eq.([33 ] ) for the tree amplitude so that the effective final state phase @xmath164 the phase @xmath96 is generated by rescattering correction and its value is -8@xmath165 for the direct @xmath2 asymmetries , the relevant phases are @xmath166 and @xmath167 .
for the penguin amplitude , we assume that the effective final state phase @xmath168 has the value near the upper bound .
thus we have @xmath169 @xmath170 now @xcite @xmath171 neglecting the terms of order @xmath172 , we have @xmath173for @xmath161@xmath174now the experimental values of @xmath175 , @xmath176 and @xmath177 are @xcite@xmath178where the numerical values in the bracket are the latest experimental values as given in ref @xcite . with @xmath169
we get from eq.([16h ] ) , @xmath179 however for @xmath180 which one gets from eq.([15 ] ) for @xmath181we obtain the following values for @xmath182 from eqs.([16h ] ) and ( [ 16i])@xmath183 we conclude : the phase shift @xmath184 for @xmath185 is compatible with experimental value of the direct @xmath186asymmetry for @xmath83 decay mode . for @xmath187
is compatible with the value ( @xmath188)@xmath189 obtained by the authors of ref.[7 ] .
finally we note that the actual value of the effective phase shift ( @xmath190 depends on one free parameter @xmath191 , factorization implies @xmath192 i.e. @xmath193 for the tree amplitude ; for the penguin amplitude , @xmath194 depends on @xmath195 however , from the experimental values of the direct @xmath2-violation for @xmath196 @xmath197 , it is near the upper bound .
finally we note that @xmath198 decays are @xmath199-wave decay whereas @xmath200 decays are @xmath201wave decays . for @xmath201wave , the decay amplitude @xmath202 ( s / s_{0 } ) \\ & & + \frac{2\gamma _ { 0}^{2}}{16\pi } i\left [ \frac{1}{\ln ( s / s_{0})-i\pi } -% \frac{2}{s}\frac{1}{\left [ \ln ( s / s_{0})-i\pi \right ] ^{2}}(s / s_{0})^{-1/2}% \right ] \\ & \approx & \frac{1}{16\pi s}ic_{p}\frac{1}{b}(s / s_{0})+\frac{2\gamma _ { 0}^{2}% } { 16\pi } i\frac{1}{\ln ( s / s_{0})-i\pi } ( s / s_{0})^{-1/2}+o\left ( \frac{1}{s}% \right)\end{aligned}\]]to be compared with eq.([31a ] ) .
now for the @xmath203 decay , only longitudinal polarization of @xmath191 is effectively involved .
since the longitudinal @xmath191-meson emulates a pseudoscalar meson and if we assume same couplings as for pions , we conclude that the final state phase for @xmath204 should be of the order @xmath205 ; in any case it should not be greater than @xmath205 .
the upper bound @xmath206 can be used to select the several possible solutions in table-2 [ section-4 ] obtained from the analysis of weak decays @xmath207 .
in this section , we discuss the experimental tests to verify the equality ( implied by c - invariance of s - matrix ) of phase shifts @xmath10 and @xmath208 for the weak decays of b mesons mentioned in section 1.it is convenient to write the time - dependent decay rates in the form 13,6 @xmath209 + \left [ \gamma _ { % \bar{f}}-\bar{\gamma}_{f}(t)\right ] \notag \\ & = & e^{-\gamma t}\left\ { \cos \delta mt\left [ \left ( \left| a_{f}\right| ^{2}-\left| \bar{a}_{\bar{f}}\right| ^{2}\right ) + \left ( \left| a_{\bar{f}% } \right| ^{2}-\left| \bar{a}_{f}\right| ^{2}\right ) \right ] \right .
\notag \\ & & \left . + 2\sin \delta mt\left [ \text{im}\left ( e^{2i\phi _ { m}}a_{f}^{\ast } % \bar{a}_{f}\right ) + \text{im}\left ( e^{2i\phi _ { m}}a_{\bar{f}}^{\ast } \bar{a}% _ { \bar{f}}\right ) \right ] \right\ } \notag \\ & & \label{e1 } \\ & & \left [ \gamma _ { f}(t)+\bar{\gamma}_{\bar{f}}(t)\right ] -\left [ \gamma _ { % \bar{f}}(t)+\bar{\gamma}_{f}(t)\right ] \notag \\ & = & e^{-\gamma t}\left\ { \cos \delta mt\left [ \left ( \left| a_{f}\right| ^{2}+\left| \bar{a}_{\bar{f}}\right| ^{2}\right ) -\left ( \left| a_{\bar{f}% } \right| ^{2}+\left| \bar{a}_{f}\right| ^{2}\right ) \right ] \right .
\notag \\ & & \left .
+ 2\sin \delta mt\left [ \text{im}\left ( e^{2i\phi _ { m}}a_{f}^{\ast } % \bar{a}_{f}\right ) -\text{im}\left
( e^{2i\phi _ { m}}a_{\bar{f}}^{\ast } \bar{a}% _ { \bar{f}}\right ) \right ] \right\ } \notag \\ & & \label{e2}\end{aligned}\ ] ] * case ( i ) : * eqs .
@xmath210 and @xmath211 give @xmath212+[\gamma _
{ \bar{f}}(t)-\bar{\gamma}_{f}(t)]}{[\gamma _ { f}(t)+\bar{% \gamma}_{\bar{f}}(t)]+[\gamma _ { \bar{f}}(t)+\bar{\gamma}_{f } ] } \notag \\ & = & \frac{2\bigl|f_{f}\bigr|\bigl|f_{\bar{f}}^{^{\prime } } \bigr|}{\bigl|f_{f}% \bigr|^{2}+\bigl|f_{\bar{f}}^{^{\prime } } \bigr|^{2}}\sin \delta mt\sin \bigl(% 2\phi _ { m}-\phi -\phi ^{^{\prime } } \bigr)\cos \bigl(\delta _ { f}-\delta _ { % \bar{f}}^{^{\prime } } \bigr ) \label{e6 } \\ \mathcal{f}\left ( t\right ) & \equiv & \frac{\left [ \gamma _ { f}(t)+\bar{\gamma}% _ { \bar{f}}\right ] -\left [ \gamma _ { \bar{f}}(t)+\bar{\gamma}_{f}\right ] } { % \left [ \gamma _ { f}(t)+\bar{\gamma}_{\bar{f}}\right ] + \left [ \gamma _ { \bar{f}% } ( t)+\bar{\gamma}_{{f}}\right ] } \notag \\ & = & \frac{\bigl|f_{f}\bigr|^{2}-\bigl|f_{\bar{f}}^{^{\prime } } \bigr|^{2}}{% \bigl|f_{f}\bigr|^{2}+\bigl|f_{\bar{f}}^{^{\prime } } \bigr|^{2}}\cos \delta mt \notag \\ & -&\frac{2\bigl|f_{f}\bigr|\bigl|f_{\bar{f}}^{^{\prime } } \bigr|}{\bigl|f_{f}% \bigr|^{2}+\bigl|f_{\bar{f}}^{^{\prime } } \bigr|^{2}}\sin \delta mt\cos \left ( 2\phi _ { m}-\phi -\phi ^{^{\prime } } \right ) \sin \bigl(\delta _ { f}-\delta _ { \bar{f}}^{^{\prime } } \bigr ) \label{e7}\end{aligned}\ ] ] the effective lagrangians @xmath50 and @xmath213 are given by @xmath214 @xmath215 [ \bar{c } \gamma_{\mu } ( 1 - \gamma_{5 } ) b ] \notag \\
\mathcal{l}_{w}^{^{\prime } } & = v_{ub } v_{cq}^{\ast } [ \bar{q } \gamma^{\mu } ( 1 - \gamma^{5 } ) c ] [ \bar{u } \gamma_{\mu } ( 1 - \gamma_{5 } ) b ] \label{cc1 } \\\end{aligned}\ ] ] we note that for time integrated @xmath2-asymmetry , @xmath227 dt}{\int_{0}^{\infty } % \left [ \gamma _ { fs}\left ( t\right ) + \bar{\gamma}_{fs}\left ( t\right ) \right ] dt } \notag \\ = & -\frac{2r_{d_{s}}r}{1+r_{d_{s}}^{2}}\sin \left ( 2\beta _ { s}+\gamma \right ) \frac{\delta m_{b_{s}}/\gamma _ { s}}{1+\left ( \delta m_{b_{s}}/\gamma _ { s}\right ) ^{2}}\cos ( \delta _ { f_{s}}-\delta _ { \bar{f}_{s}}^{^{\prime } } ) \label{4.35a}\end{aligned}\ ] ] for @xmath230 , replace @xmath231 , @xmath232 , @xmath233 , @xmath234 in eq . @xmath235 .
since for @xmath236 , in the standard model , with three generations , gives @xmath237 , so we have for the cp - asymmetries @xmath238 or @xmath239 instead of @xmath240 , @xmath241 .
hence @xmath236-decays are more suitable for testing the equality of phase shifts @xmath242 and @xmath243 as for this case neither @xmath244 nor @xmath245 is suppressed as compared to the corresponding quantities for @xmath220 . to conclude , for @xmath246 decays , the equality of phases @xmath247 and @xmath248 for @xmath249 gives @xmath250 these examples have been selected keeping in view that final state phases @xmath268 are not too large . for @xmath269 ,
we have used eq . @xmath270 neglecting the second order term .
an attractive option is @xmath271 for each value of @xmath272 ; although @xmath273 is also a possibility .
@xmath274 implies @xmath275 .
finally the cp asymmetries in the limit @xmath284 @xmath285}{\gamma ( 1+a_{cp } ) } \notag \\ & = \sqrt{1-c_{\bar{f}}^{2}}\sin ( 2\alpha _ { eff}^{\bar{f}}+\delta ) \notag \\ & = -\sqrt{1-c_{\bar{f}}^{2}}\cos \delta \label{cccc7 } \\
s_{f}=s-\delta s & = \frac{2\text{im}[e^{2i\phi _ { m}}a_{f}^{\ast } \bar{a}_{f}]% } { \gamma ( 1-a_{cp } ) } \notag \\ & = \sqrt{1-c_{f}^{2}}\sin ( 2\alpha _ { eff}^{f}-\delta ) \notag \\ & = \sqrt{1-c_{f}^{2}}\cos \delta \label{cccc8}\end{aligned}\ ] ] the final state strong phases essentially arise in terms of @xmath68-matrix , which converts an @xmath288 state into an @xmath289 state .
the isospin , @xmath1-invariance of hadronic dynamics and the unitarity together with two particle scattering amplitudes in terms of regge trajectories are used to get information about these phases .
in particular two body unitarity is used to calculate the final state phase @xmath96 generated by rescattering for the color suppressed decays in terms of the color favored decays . in the inclusive version of unitarity
, the information obtained for @xmath199-wave scattering from regge trajectories is used to derive the bounds on the final state phases .
in particular , the value obtained for the final state phases @xmath290 @xmath291 and @xmath292 is found to be compatible with the experimental values for direct @xmath2 asymmetries @xmath293 . for @xmath294 ,
@xmath295 decays described by two independent single amplitudes @xmath6 , @xmath296 and @xmath297 @xmath298 with different weak phases viz . @xmath8 and @xmath9 , equality of phases @xmath299 implies , the time dependent cp asymmetries @xmath300 an added advantage is that these decays are described by tree graphs .
assuming factorization , the decay amplitude @xmath6 can be determined in term of the form factors @xmath301 and @xmath302 .
the parameter @xmath303 can be expressed in terms of the ratios of the form factors @xmath304/@xmath305 and @xmath306/@xmath307 . from the experimental branching ratios ,
we have obtained the form factors @xmath301 and @xmath308 which are in excellent agreement with the prediction of hqet .
we have also determined @xmath309 .
for @xmath309 we get the value @xmath310 . using this value
we get the following bound from the experimental value of @xmath311 for @xmath312 decay : @xmath313 in section-4 , the decays @xmath316 for which decay amplitudes @xmath21 and @xmath6 are given in terms of tree and penguin diagrams are discussed .
we have analyzed these decays assuming factorization for the tree graph .
factorization implies @xmath317 . in the limit @xmath318 , we have shown that @xmath319 the first equation has been solved graphically , from which the final state phases @xmath320 corresponding to various values of @xmath321 can be found for a particular value of @xmath322 .
the upper bound @xmath323 obtained in section-2 , using unitarity and strong interaction dynamics based on regge pole phenomonalogy can be used to select the solutions given in table-2 . neglecting the terms of order @xmath324 , we get using factorization @xmath325 l. wolfenstein , hep - ph/0407344 v1 ( 2004 ) ; n. spokvich , nuovo cinento , 26 , 186 ( 1962 ) ; k. gottfried and j. d. jackson , nuovo einento 34 , 735 ( 1964 ) ; see also , fayyazuddin and riazuddin , quantum mechanics , page 140 , world scientific ( 1990 ) . for a review ,
see for example cp - violation editor : c. jarlskog , world scientific ( 1989 ) ; h. quinn , b physics and cp - violation [ hep - ph/0111174 ] .
fayyazuddin and riazuddin , a modern introduction to particle physics , 2nd edition , world scientific .
plot of equation @xmath328 for different values of @xmath329 for @xmath330 where solid curve , dashed curve , dashed doted curve , dashed bouble doted and double dashed doted curve are corresponding to @xmath331 and @xmath332 respectively . | using the unitarity , @xmath0 and @xmath1-invariance of hadronic interactions , the bounds on final state phases are derived .
it is shown that values obtained for the final state phases relevant for the direct @xmath2-asymmetries @xmath3 are compatiable with experimental values for these asymmetries .
for the decays @xmath4 @xmath5 described by two independent single amplitudes @xmath6 and @xmath7 with differnt weak phases ( @xmath8 and @xmath9 ) it is argued that the @xmath1-invariance of hadronic interactions implies the equality of the final state phase @xmath10 and @xmath11 .
this in turn implies , the @xmath2-asymmetry @xmath12 is determined by weak phase ( @xmath13 only whereas @xmath14 assuming factorization for tree graphs , it is shown that the @xmath15 form factors are in excellent agreement with heavy quark effective theory . from the experimental value for @xmath16 the bound @xmath17
is obtained and @xmath18 is predicted . for the decays described by the amplitudes @xmath19 such as @xmath20 @xmath21 and @xmath22 where these amplitudes are given by tree and penguin diagrams with
differnt weak phases , it is shown that in the limit @xmath23 and @xmath24 |
ground - based and spaced - based observations have allowed many supernova remnants ( snrs ) in our galaxy to be resolved , showing various different morphologies in radio , x - ray , and optical wavelengths ( weiler & sramek 1988 ; van den bergh 1988 ; rho & petre 1998 ) . the expanding snr shock can heat the surrounding interstellar medium ( ism ) to temperatures up to @xmath0 k and beyond , producing thermal x - ray emission , while synchrotron radiation seen in the radio band is due to the relativistic electrons accelerated behind the shock front in the magnetic field of the snr .
optical emission from snrs is from shocked - heated collisionally ionized species such as [ s ii ] , [ n ii ] and [ o iii ] , as well as h@xmath1 recombination emission .
multiwavelength observations of snrs therefore provide important information on the structure of the ism and chemical composition in the host galaxy , the evolutionary stage of the snr , the character of the supernova ejecta and the nature of the progenitor and its associated stellar winds .
most detailed studies of snr have concentrated on our galaxy simply because of the large angular size of the snrs .
however , studies of galactic snrs can be limited by the lack of reliable distance estimates , high interstellar absorption at optical and x - ray wavelengths , and the high level of confusion in many galactic fields ( magnier et al .
these limitations are overcome by studies of snrs in nearby extragalactic systems such as the large and small magellanic clouds ( e.g. , williams et al . 1999 ) , and the andromeda galaxy ( m31 ) . at a distance of 780 kpc ( stanek & garnavich 1998 )
, m31 is the closest spiral galaxy that shares similar morphology , metallicity and size with the milky way ; it is the best candidate for a comparative study of snrs .
studies of snrs in m31 are mainly done in the optical ( e.g. , dodorico , dopita , & benvenuti 1980 ; blair , kirshner , & chevalier 1981 , 1982 ; braun & walterbos 1993 ; magnier et al . 1995 ) .
about 200 snr candidates have been identified by these surveys . in a recent extensive _ rosat _ position sensitive proportional counter ( pspc ) survey ( supper et al .
2001 ) of m31 , 16 x - ray snrs were identified within a total area of @xmath2 deg@xmath3 by cross - correlating with these optical catalogs .
their x - ray luminosities ( 0.12.4 kev ) range from @xmath4 to @xmath5 .
more recently , kong et al . (
2002a ) discovered a spatially resolved x - ray snr in m31 with _
chandra_. in the radio , sjouwerman & dickel ( 2001 ) found 3 snrs near the center of m31 . in this _
letter _ , we report the discovery of the first radio / x - ray / optical resolved snrs in the center of m31 .
at the distance of m31 , the typical diameter of snrs is between @xmath6 arcsec and 20 arcsec so that some of the snrs are large enough to be studied with high - resolution instruments in sufficient detail ( sjouwerman & dickel 2001 ; kong et al .
in particular , sensitive interferometric radio images have sufficient resolution to resolve the snrs without suffering from the large scale diffuse background in the center of m31 ( hjellming & smarr 1982 ; crane , dickel , & cowan 1992 ; crane et al .
1993 ) . very sensitive 8.4 ghz ( 3.6 cm ) radio observations of the central 500 arcsec ( 1.9 kpc ) of m31 were obtained by combining in total 120 hours of archival data taken with the very large array ( vla ) in its largest array configurations ( a and b ) between 1990 and 1996 .
the resulting radio image has an angular resolution of 0.24 arcsec and is sensitive to angular sizes of up to 20 arcsec , and has an rms noise level of @xmath7 4 @xmath8jy / beam . after discovering three faint radio snrs in the 8.4 ghz archival data
, we also observed the remnants for 24 hours with the vla at 4.9 ghz ( 6 cm ) in july 2002 . the recent 4.9 ghz radio image ( fig .
1 ) is sensitive to angular structures from 1.2 to 36 arcsec and has an rms of @xmath7 5 @xmath8jy / beam .
both radio images clearly show resolved shell - like objects that are brighter at 4.9 ghz than at 8.4 ghz ( sjouwerman & dickel 2001 )
. the largest snr is located at 70 arcsec east - south - east of the nucleus of m31 at a position of r.a.=00@xmath942@[email protected] , dec.=+41@xmath1215@[email protected] ( j2000 ) , labeled object 1 in fig .
the integrated flux density of the complex radio morphology is measured to be 1.13 mjy at 4.9 ghz .
its size of @xmath15 arcsec ( @xmath16 pc ) is consistent with a previous @xmath17 arcsec low - resolution ( 5 arcsec ) vla detection at 1.4 ghz designated braun-101 ( braun 1990 ) . with the notion that the flux density measured at 1.4 ghz might be confused with the extended background ( hjellming & smarr 1982 ; braun 1990 ) , we derive a spectral index between 1.4 and 4.9 ghz of @xmath1 = @xmath180.77 or flatter , where @xmath19 .
the radio morphology and radio spectral index are consistent with a shell - type snr ( dickel & milne 1998 ) .
it is worth noting that our radio observations also discover two more new snrs , designated braun-80 and braun-95 ( braun 1990 ) , near the nucleus of m31 ( sjouwerman & dickel 2001 ) . following the discovery of the radio snrs , we compare their radio positions with the m31 x - ray source positions as detected by the _ chandra _
x - ray observatory ( kong et al . 2002b ; kaaret 2002 ) and indeed the largest radio snr , braun-101 , is spatially coincident with the x - ray source cxom31 j004250.5 + 411556 ( kong et al .
we further examine a 40 ksec ( 37.7 ksec after rejecting high background period ) _ chandra_x - ray image taken with the back - illuminated chip ( s3 ) of the advanced ccd imaging spectrometer ( acis ) .
the snr is clearly detected in the 0.37 kev image and it is resolved as a ring - like object with a diameter of about 5 arcsec ( fig .
the snr has about 55 background - subtracted photons ( 0.37 kev ) in a @xmath20 radius extraction region .
all of the emission from the snr is confined to energies below @xmath6 kev , indicating that it is a soft x - ray source ( e.g. , distefano & kong 2003 ) .
the shell - like x - ray morphology , soft x - ray nature , and good positional coincidence with the radio snr makes the x - ray source as a secure x - ray counterpart of the radio identified snr .
we extracted the energy spectrum from a @xmath20 radius and background from an annulus centered on the source . in order to allow @xmath21 statistics to be used
, the spectrum was grouped into at least 10 counts per spectral bin .
we fitted the data with a raymond - smith model and kept the @xmath22 fixed at @xmath23 @xmath24 .
the best fit temperature is @xmath25 kev ( 90% confidence ) with a value of reduced chi - squared statistic @xmath26 for 5 degrees of freedom ( d.o.f . ) .
the 0.37 kev luminosity of the best fitting spectrum is @xmath27 .
we also re - examined the 50 ksec _
_ high resolution camera ( hrc ) observation ( kaaret 2002 ) and braun-101 is indeed visible as a faint extended source though it was not in the published hrc source list . assuming the above spectral model , the corresponding 0.37 kev luminosity of the snr is @xmath28 .
the other radio snr , braun-95 , is also detected in the _ chandra _ observation as an extended ( with a diameter of @xmath29 arcsec ) object ( marked as 2 in fig .
although it shows a clear shell morphology in the radio image , the x - ray emission is very patchy .
we extracted the energy spectrum with a @xmath30 radius circle centered on the source and background from a source - free region .
we kept the @xmath22 fixed at @xmath23 @xmath24 and fitted the spectrum with raymond - smith model .
the fit is rather poor ( @xmath31 for 10 d.o.f . ) .
we then added an additional power - law component and the fit was improved ( @xmath32 for 8 d.o.f . ) .
the best fit temperature is @xmath33 kev , while the photon index is @xmath34 ; the 0.37 kev luminosity is @xmath35 . with the detection of radio and x - ray emissions from braun-101 , we searched the _ hubble space telescope _ ( _ hst _ ) archive for any narrow - band exposures near that position .
we found images taken with the wide field planetary camera 2 ( wfpc2 ) in h@xmath1 ( f656n ) filter .
we downloaded the pipeline - calibrated images , registered them , and combined them with cosmic - ray rejection . the resulting continuum ( f547 m filter ) subtracted image of the region around the snr
is shown in figure 2 . a resolved shell - like object with 3 arcsec diameter
is seen at the radio and x - ray position in the h@xmath1 image , indicating an optical counterpart of the snr .
the total h@xmath1 flux is @xmath36 erg @xmath24 s@xmath37 , while the peak h@xmath1 surface brightness of the snr is @xmath38 erg @xmath24 s@xmath39 arcsec@xmath40 in the continuum - subtracted h@xmath1 image .
re - examination of ground - based narrow - band optical images from the kitt peak national observatory 0.6-m burrell schmidt telescope also reveals braun-101 ( but unresolved ) with [ s ii]/h@xmath41 .
this high ratio makes it unlikely that the h@xmath1 emission is from a hii region near the snr ( levenson et al . 1995 ) , and confirms the radio / x - ray identification as a snr .
unfortunately , radio / x - ray snr , braun-95 , is just outside the field of view of the _ hst _ observations .
using vla , _ chandra _ , and _ hst _ data , we found one radio / x - ray / optical resolved snr ( braun-101 ) and one radio / x - ray resolved snr ( braun-95 ) near the center of m31 .
the radio emission of braun-101 traces the outer shell of the remnant ; it is an asymmetric shell with the northeastern side about three times brighter ( 67 versus 23 @xmath8jy / beam ) than the southwestern side .
the x - ray remnant with a diameter of about 5 arcsec ( 20 pc ) is considerably smaller than the radio remnant and the x - ray outline lies predominantly inside the brighter part of the radio remnant .
both wavelengths show a shell configuration but the x - ray remnant is elongated toward the north , while the southern part of the x - ray remnant near the center of the radio remnant is slightly brighter in x - ray .
it could be that the situation is similar to that seen for snr 1e0102.2 - 7219 in the small magellanic cloud ( amy & ball 1999 ; gaetz et al . 2000 ) .
the x - ray emission is brightest in a ring primarily inside the radio shell but very faint x - rays are seen extending out to the shock just outside the radio shell .
the bright inner x - ray ring is interpreted as coming from heating and ionization from the reverse shock moving into a strong gradient in pre - explosion ejecta of the progenitor star . a longer x - ray exposure might reveal the expected faint outer x - ray emission .
the larger radio shell is perhaps the forefront of the shock in a relatively low ism density toward southwest , where the greatest magnetic field compression and particle acceleration occurs .
the optical emission of braun-101 is the smallest ( 3 arcsec ; 11 pc ) compared to other wavelengths and it is encompassed by the x - ray emission ( fig .
the h@xmath1 shell roughly traces the inner x - ray ring of the remnant and shows a non - uniform distribution of emission over the shell , with slight enhancement toward the southwest .
it also shows a crescent opening and possibly a double - ring structure in the northern region of the optical remnant . in the northeast , a local enhancement of
the x - ray emission corresponds to a local optical enhancement
. the optical emission could be produced by shocks driven into cooler and denser material by the high pressure downstream of the reverse shock .
it is worth noting that such large discrepancies in snr size at different wavelengths are not typical ( winkler & long 1997 ) . in this case
, we may be seeing only the brightest optical knots and not the whole optical shell .
braun-101 is only 0.5 arcsec from an optically identified planetary nebula candidate ( kong et al .
2002b ; ciardullo et al .
our multiwavelength observations provide strong evidence that the source is actually a snr . as planetary nebulae
are usually identified by their strong [ o iii ] emission ( ciardullo et al .
1989 ) , they could resemble the [ o iii ] emission from some snrs .
however , the large angular size ( from 10 pc in the optical to 40 pc in the radio ) and high x - ray luminosity ( @xmath42 erg s@xmath39 ) rule out the possibility of a planetary nebula for which the typical diameter and x - ray luminosity is @xmath43 pc ( @xmath44 arcsec ) and @xmath45 erg s@xmath39 , respectively . like braun-101 ,
braun-95 is clearly a shell - like radio remnant ( see fig .
1 ) and its morphology is also very similar to braun-101 .
the x - ray remnant of braun-95 , however , is patchy and is mainly located in the northeastern side of the radio remnant , where the radio emission is brightest .
faint x - ray emission is also seen near the radio shell .
the x - ray spectrum likely consists of two components : a raymond - smith model and a power - law model .
the raymond - smith temperature is similar to braun-101 and cxom31 j004327.7 + 411829 ( kong et al .
the extra power - law component might hint that braun-95 contains a pulsar and/or a pulsar nebula .
discovery of snrs near the center of m31 provides some information about the star formation history of the region .
for instance , type ia snrs suggest an old stellar population , while type ib / ii snrs are associated with young population .
unfortunately , we do not have direct evidence about the type for the two snrs based on the current multi - wavelength data . by comparing with galactic snrs , the mismatch in morphology of the two snrs in different wavelengths
might hint the type .
for example , ic 443 ( petre et al .
1988 ; kawasaki et al .
2002 ) and sgr a east ( maeda et al .
2002 ) are type ii snrs that have different morphology in radio and x - ray , and type ia snr like tycho and sn 1006 generally have a well matched morphology in x - ray and radio . on the other hand , the x - ray and radio surface brightness of tycho does not show correspondence ( hwang & gotthelf 1997 ) , and the x - ray emission of sn 1006 is non - thermal ( long et al .
2003 ) , suggesting a synchrotron origin of x - rays .
in addition , type ib / ii snrs such as cas a ( compare hughes , et al . 2000 and braun , gull & perley 1987 ; see also keohane et al .
1996 and dickel et al . 1982 ) and rcw103 ( compare dickel et al .
1996 and tuohy & garmire 1980 ) also have a good match in different wavelengths .
further multi - wavelength observaitons of these 2 snrs are required to identify their type and hence the assoicated stellar population in the bulge of m31 .
the multiwavelength detection of snrs in m31 can also provide further information about the local ism density around the snrs .
the ism near m31 s snrs is thought to be low ( @xmath46 @xmath47 ) , which is 1 to 2 orders of magnitude smaller than the measurements by the hi observations ( magnier et al .
our discovery of x - ray emission from snrs near the nucleus of m31 indicates that the ism density in the vicinity of the snrs could be higher .
we have shown that by combining all the high - resolution instruments across many wavebands , we can study the morphological structure of snrs in m31 for the first time .
this is a first step toward understanding the evolutionary states and shock morphologies of snrs in m31 .
more specific multiwavelength observations in the future will shed more light on the properties of snrs in m31 , allowing more comprehensive comparative studies with our own galaxy .
amy , s. , & ball , l. , 1999 , apj , 411 , 761 blair , w.p . ,
kirshner , r.p .
, & chevalier , r.a .
1981 , apj , 247 , 879 blair , w.p . ,
kirshner , r.p .
, & chevalier , r.a .
1982 , apj , 254 , 50 braun , r. , & walterbos , r.a.m .
1993 , a&as , 98 , 327 braun , r. 1990 , apjs , 72 , 761 braun , r. , gull , s.f . , perley , r.a .
1987 , nature , 327 , 395 ciardullo , r. , jacoby , g.h . ,
ford , h.c . , & neill , j.d . , 1989 , apj , 339 , 53 crane , p.c . , dickel j.r . ,
& cowan , j.j . , 1992 , apj , 390 , l9 crane , p.c . , cowan , j.j . ,
dickel , j.r .
, & roberts , d.a . , 1993 ,
apj , 417 , l61 dickel , j.r . , & and milne , d.k . , 1998 ,
aj , 115 , 1057 dickel , j.r . , green , a. , ye , t. , milne , d.k .
1996 , aj , 111 , 340 dickel , j.r . ,
murray , s.s . , morris , j. , wells , d.c .
1982 , apj , 257 , 145 distefano , r. , & kong , a.k.h . , 2003 ,
apj , in press ( astro - ph/0301162 ) dodorico , s. , dopita , m.a . , & benvenuti , p. 1980 , a&as , 40 , 67 gaetz , t.j . , butt , y.m . , edgar , r.j . , eriksen , k.a . ,
plucinsky , p.p . ,
schlegel , e.m . , & smith , r.k . , 2000 , apj , 543 , l47 hjellming , r.m .
, & smarr , l.l . , 1982 , apj , 257 , l13 hughes , j.p . ,
rakowski , c.e . , burrows , d.n . , & slane , p.o .
2000 , apj , 528 , l109 hwang , u. , & gotthelf , e.v .
1997 , apj , 475 , 665 kaaret , p. , 2002 ,
apj , 578 , 114 kawasaki , t.k . , ozaki , m. , nagase , f. , masai , k. , ishida , m. , & petre , r. 2002 , apj , 572 , 897 keohane , j.w . ,
rudnick , l. , & anderson , m.c .
1996 , apj , 466 , 309 kong , a.k.h . ,
garcia , m.r .
, primini , f.a . , & murray , s.s .
, 2002a , apj , 580 , l125 .
kong , a.k.h . ,
garcia , m.r .
, primini , f.a . ,
murray , s.s . ,
distefano , r. , & mcclintock , j.e . , 2002b , apj , 577 , 738 lauer , t.r .
, et al . , 1993 , aj , 106 , 1436 levenson , n.a . , krishner , r.p . , blair , w.p . , & winkler , p.f . , 1995 ,
aj , 110 , 739 long , k.s .
, reynolds , s.p . ,
raymond , j.c . ,
winkler , p.f .
, dyer , k.k . , & petre , r. 2003 , apj , 586 , 1162 maeda , y. , et al .
2002 , apj , 570 , 671 magnier , e.a .
, prins , s. , van paradijs , j. , lewin , w.h.g . , supper , r. , hasinger , g. , pietsch , w. , & truemper , j. , 1995 , a&as , 114 , 215 magnier , e.a . , primini , f.a .
prins , s. , van paradijs , j. , & lewin , w.h.g . , 1997 , apj , 490 , 649 petre , r. , szymkowiak , a.e . , seward , f.d . , & willingale , r. 1988 , apj , 335 , 215 rho , j. , & petre , r. , 1998 , apj , 503 , l167 stanek , k.z . , & garnavich , p.m. , 1998 , apj , 503 , l131 supper , r. , hasinger , g. , lewin , w.h.g .
, magnier , e.a . ,
van paradijs , j. , pietsch , w. , read , a.m. , & trumper , j. 2001 , a&a , 373 , 63 sjouwerman , l.o . , & dickel , j.r . , 2001 , in _ young supernova remnants _ , eds .
s.s . holt and u. hwang , aip conference proceedings , vol .
565 , p.433 tuohy , i. , & garmire , g. 1980 , apj , 239 , l107 van den bergh , s. , 1988 , apj , 327 , 156 weiler , k.w . , & sramek , r.a . , 1988 ,
ara&a , 26 , 295 williams , r.m , chu , y. , dickel , j.r .
, petre , r. , smith , r.c . , & tavarez , m. , 1999 , apjs , 123 , 467 winkler , p.f .
, & long , k.s . , 1997 , apj , 491 , 829 | we have detected a spatially resolved supernova remnant ( snr ) in the center of the andromeda galaxy , in radio , x - ray , and optical wavelengths .
these observations provide the highest spatial resolution imaging of a radio / x - ray / optical snr in that galaxy to date . the multi - wavelength morphology , radio spectral index , x - ray colors , and narrow - band optical imaging are consistent with a shell - type snr .
a second snr is also seen resolved in both radio and x - ray . by comparing the morphological sturcture of the snrs in different wavelengths and with that in our own galaxy , we can study the shock morphologies of snrs in the andromeda galaxy .
the proximity of the snrs to the core suggests high interstellar medium density in the vicinity of the snrs in the center of the andromeda galaxy . #
1to 0pt#1 [ firstpage ] |
chiral spin textures are studied now very actively for possible spin self - organization , unusual quantum transport phenomena and spintronic applications .
a well established mechanism of spin chirality is the spin - orbit dzyaloshinskii
moriya ( dm ) interaction which is responsible for intricate magnetic patterns in the mnsi - type crystals .
even half a century after the discovery of the strange magnetic properties of mnsi @xcite , the magnetics with the @xmath0 crystal structure still amaze us with the variety and complexity of their magnetic phases and electronic properties @xcite contrasting the simplicity of the crystalline arrangement ( only four magnetic atoms per a unit cell ) . among the magnetic phases ,
both experimentally observed and hypothetical , are simple and cone helices @xcite , the skyrmions and their lattices associated with the recently found @xmath3-phase @xcite , possible 3d structures @xcite similar to the blue phases of liquid crystals , etc .
this variety is due to , first , the lack of inverse and mirror symmetries , which gives rise to the chirality of the crystalline and spin structures ; second , the frustrations @xcite resulting from nontrivial topology of the trillium lattice , that introduces a competition of various interactions between different pairs of atoms .
beginning from the discovery of the chiral magnetic properties of mnsi in 1976 @xcite till present day the most used approach to describe and predict twisted magnetic structures remains the phenomenological theory based on the ginsburg landau free energy with an additional term first introduced by dzyaloshinskii @xcite .
however , the approach , which uses our knowledge of the system symmetry , is not able to say anything about the values of coefficients in the free energy , for instance , how they are connected with the real interactions between atoms .
the microscopic theories , e.g. the model of the classical heisenberg ferromagnetics with a spin - orbit term originally developed by moriya @xcite , have in their turn the shortcoming that the number of variables , including the spins of all the magnetic atoms , is infinite , and therefore they are difficult to use for any analytical computations .
nevertheless , in spite of some doubts about validity the heisenberg model in the itinerant magnetics , it is often used for digital simulations @xcite . for this model
it is of great importance to know the parameters of different interactions between pairs of spins : @xmath4 for exchange coupling of @xmath5th and @xmath6th atoms , and vectors @xmath7 for dm interactions .
those parameters can be obtained in two ways : from comparison of theoretical results with an experimental data and by means of _ ab intio _ calculations . in our recent work @xcite , using a coarse - grained approximation , the phenomenological constants @xmath8 and @xmath9 of the elastic free energy have been connected with the corresponding parameters of the microscopic theory .
a surprising detail was the recognition of inter - sublattice canting as a new microscopic feature of the magnetic structures in the mnsi - type crystals .
[ fig - canting ] shows the difference between the twist and the canting by example of a 1d spin chain .
an experimental confirmation of the canting could give us an argument for applicability of the heisenberg model to the itinerant ferromagnetics of mnsi - type @xcite .
similar features could be observed in dielectric chiral magnetics like recently studied multiferroics cu@xmath10oseo@xmath11 @xcite , bifeo@xmath11 @xcite and langasite - type crystals @xcite .
though the theory with only nearest neighbors interactions is frequently used to simulate spin structures @xcite , it does not always give an adequate physical description .
indeed , very often the coupling between second and next neighbors is comparable or even more considerable than that of the nearest neighbors .
thus , for example , in the weak ferromagnetic @xmath12-fe@xmath10o@xmath11 , the antiferromagnetic exchange interactions and dm interactions are expected to be the most strong for third and fourth neighbors ( see _ ab initio _ calculations in @xcite ) ; this is also supported by the experimental data for the heisenberg exchange parameters @xcite .
it also could happen that only with the interactions between non - nearest neighbors one can describe an experimentally observed phenomenon .
for example , if in a 1d spin chain with the ferromagnetic coupling between the nearest neighbors ( @xmath13 ) one turns on the antiferromagnetic interaction with the second neighbors ( @xmath14 ) , then for @xmath15 a spiral magnetic structure appears owing to spontaneous breaking of the inverse symmetry .
such approach can explain the phase transition appearing at @xmath1628 k and inducing ferroelectricity in the multiferroic tb(dy)mno@xmath11 crystal @xcite .
the approach works also in the itinerant magnetics of mnsi - type .
just the competition between ferromagnetic coupling of the nearest spins ( @xmath17 ) and antiferromagnetic coupling of the second and third neighbors ( @xmath18 ) was utilized to explain the observed alignment of magnetic helices along crystallographic directions @xmath19 in mnsi at high pressure @xcite . in this paper
, we suggest a possible mechanism for effective reinforcement of the twist terms in the chiral spin structures of the @xmath0 magnetics , taking into account the interactions with non - nearest neighbors .
the reinforcement is caused by an interplay between the exchange coupling with the second and third neighbors and the canting from dm interactions . in sections
[ sec : interactions][sec : energy ] the transition from the heisenberg microscopic model to the continuous phenomenological one is performed , and the spin - orbit terms in the phenomenological energy are found up to the contributions of the order of @xmath20 . in sec .
[ sec : extratwist ] the possibility of an extra twist induced by the canting is demonstrated . in sec .
[ sec : rkky ] the exchange interaction parameters are estimated within the rkky theory . in sec . [
sec : experimental ] the possibility of an experimental proof of the canting existence is discussed .
the possible applications of the theory are suggested in sec .
[ sec : discussion ] .
both phenomenological and microscopical theories describing twisted mnsi - type magnetics contain terms induced by chirality of the system and associated with the dzyaloshinskii
moriya interaction of the spin - orbit nature . in the heisenberg model ,
the extra term being associated with an individual bond , e.g. connecting magnetic atoms 1 and 2 , can be expressed as @xmath21 , \ ] ] where @xmath22 and @xmath23 are the classical spins of the atoms , @xmath24 is the dm vector of the bond 1 - 2 . in principle , there could be magnetic moments at si atoms @xcite but this effect will be neglected below .
let us shortly describe the properties of the dm vector @xcite .
( i ) as it is obvious from eq .
( [ eq : heisenberg ] ) , the sign of the dm vector depends on which atom of the bond we consider as the first . indeed , because the cross product changes its sign when @xmath22 and @xmath23 are rearranged , then @xmath25 .
( ii ) the structure changes its chirality under inversion , whereas the energy remains unchanged , which means that @xmath24 is a pseudovector , because @xmath26 $ ] is a pseudovector as well .
( iii ) the dm vector of a bond possesses the local symmetry of the bond .
thus , in mnsi the dm vectors are of their most general form , i.e. they have three independent components , because the mn - mn bonds in the @xmath0 structure do not possess any internal symmetry element .
( iv ) the dm vectors vary from bond to bond ; the dm vectors of the equivalent bonds have the same length but may be of different orientation ( just the case of mnsi ) . in particular , if two bonds are connected by the rotation symmetry transformation of the space group @xmath27 of the crystal , then their dm vectors do by the corresponding rotation of the point group @xmath28 . in the phenomenological theory ,
the chiral interaction is induced by the following extra term @xmath29\ ] ] in the expression for the magnetic energy density . here
@xmath30 is a continuous field of the magnetic moment , @xmath9 is a pseudoscalar constant of the interaction .
because both the theories describe the same matter , there should be a relationship between them .
in particular , the terms ( [ eq : heisenberg ] ) and ( [ eq : phenomenological ] ) of the different approaches should be somehow connected .
indeed , it was shown in ref .
@xcite that in the nearest neighbors approximation the constant @xmath9 of the phenomenological theory is proportional to the component @xmath31 of the dm vector of the bond @xmath32 between neighboring manganese atoms .
however , there is a problem here .
according to different spin - orbit calculation schemes @xcite , the dm vector associated with a bond should be perpendicular or almost perpendicular to the bond . in the present case
it means that the component @xmath31 , which lies almost along the bond , constitutes only a small part of the dm vector length .
taking into account that having the relativistic nature spin - orbit dm interaction serves as a small additive to the ferromagnetic exchange coupling , we can conclude that the twist observed in the mnsi - type crystals , particularly in mnge , seems to be abnormally strong .
a possible solution of this problem is that the nearest neighbors approximation is not sufficient and hereinafter in the paper we develop this idea in details . in the following sections we will show , how to perform transformation from the microscopic heisenberg model to the macroscopic elastic theory .
in the classical heisenberg model of magnetics with an extra interaction of the dzyaloshinskii moriya type , the energy of the system is expressed as a sum of pair interactions between magnetic atoms and the interaction of individual atoms with an external magnetic field @xmath33 : @xmath34 - \frac{(\mathbf{d}_{n , ij } \cdot \mathbf{s}_i)(\mathbf{d}_{n , ij } \cdot \mathbf{s}_j)}{2j_n } \right\ } - g \mu_b \mathbf{h } \cdot \sum_{i=1}^4 \mathbf{s}_i \right ) .\ ] ] here the external summation is over all the unit cells of the crystal , @xmath35 enumerates magnetic coordination shells , the sum @xmath36 is over all the bonds of the shell @xmath35 , the sum over @xmath5 is taken over all magnetic atoms in the unit cell ; @xmath37 , @xmath38 is the exchange coupling interaction parameter of the @xmath35th shell , @xmath39 is the dm vector of the bond @xmath40 of the @xmath35th shell , @xmath41 are classical spins of magnetic atoms , @xmath42 , @xmath43 is the effective magnetic moment of the atom .
the terms of the order of @xmath44 are usually ignored but sometimes they can be important @xcite .
the mnsi - type crystal has @xmath0 structure ( @xmath27 space group ) with four magnetic ( mn ) atoms occupying crystallographically equivalent @xmath45 positions within a unit cell , @xmath46 , @xmath47 , @xmath48 , @xmath49 @xcite .
the shortest bonds between the magnetic atoms have the length @xmath50 .
the next environment consists of two close magnetic shells with the radii @xmath51 and @xmath52 .
the parameter @xmath53 defines which of them is closer to the initial atom .
when @xmath54 ( @xmath55 in the case of mnsi ) , then the shell with the radius @xmath56 is closer , so we will refer to it as the second one .
when @xmath57 , the 2nd and 3rd neighbors are at the same distance so that the manganese sublattice gains the space group @xmath58 @xcite , connecting the 2nd and 3rd shells by a symmetry transformation .
this could establish a linkage between the dm vectors of the shells , but the silicon sublattice ( having in its turn the space group @xmath59 , when @xmath60 ) breaks the symmetry and the dm vectors of the shells are different even in this case .
if @xmath61 , then the 3rd neighbors become closer than the 2nd ones .
every magnetic atom possesses six neighbors in each of its three nearest magnetic shells , so there are 36 bonds in a unit cell ( in twelves for each kind ) .
[ fig - neighbors ] shows the magnetic environment of a manganese atom in mnsi .
the atom lies on a 3-fold axis .
each of the first three magnetic shells consists of two equilateral triangles .
all the atoms of the 2nd shell are on one side of the plane perpendicular to the 3-fold axis , all the atoms of the 3rd shell are situated on the other side of the plane .
let @xmath62 be the coordinates of the vector @xmath63 corresponding to the bond @xmath64 directed from @xmath65 to @xmath66 , @xmath67 be the coordinates of the vector @xmath68 corresponding to the bond @xmath69 directed from @xmath65 to @xmath70 , and @xmath71 be the coordinates of the vector @xmath72 corresponding to the bond @xmath73 directed from @xmath65 to @xmath74 .
we specify directions for all twelve bonds of a shell in the way that all directed distances @xmath75 between neighboring atoms connect to each other by the symmetry transformations of the point group @xmath28 , so corresponding dm vectors are connected by the same transformations . in the mnsi - type crystals
there is only one type of magnetic bonds directed from a central atom of the type @xmath76 to an atom of the type @xmath77 at a given shell ( @xmath78 ) .
we can take advantage from this fact and introduce for each pair @xmath40 ( @xmath79 , @xmath80 ) a local triad @xmath81 , \ ] ] and then use it as a basis for vectors @xmath75 and @xmath7 . here
@xmath82 is a vector directed along the 3-fold axis passing through the position @xmath83 : @xmath84 then the bond vector and the dm vector of an arbitrary bond in three first magnetic shells can be written as [ eq : bij ] @xmath85 / 8 , \\ \mathbf{b}_{2,ij } & = & ( -8x+3 ) ( \boldsymbol{\tau}_i - \boldsymbol{\tau}_j ) / 8 + ( \boldsymbol{\tau}_i + \boldsymbol{\tau}_j ) / 4 - [ \boldsymbol{\tau}_i \times \boldsymbol{\tau}_j ] / 8 , \\ \mathbf{b}_{3,ij } & = & ( -8x-1 ) ( \boldsymbol{\tau}_i - \boldsymbol{\tau}_j ) / 8 + ( \boldsymbol{\tau}_i + \boldsymbol{\tau}_j ) / 4 - [ \boldsymbol{\tau}_i \times \boldsymbol{\tau}_j ] / 8 , \end{aligned}\ ] ] @xmath86 , \ ] ] with @xmath87 ; here @xmath88 because the indices @xmath40 and @xmath89 designate two different bonds .
notice that the energy ( [ eq : energy1 ] ) does not depend on the parameter @xmath53 , so it can be chosen arbitrarily , making convenient transition to the continuous approximation ( fig .
[ fig - xid](b ) ) .
indeed , to provide the transition the continuous unimodular vector functions will be defined , @xmath90 , @xmath91 , @xmath92 , @xmath93 , which coincide in special points with the spin values in corresponding atomic positions .
if the parameter @xmath53 is equal to its experimental value , e.g. @xmath94 in mnsi , then we should take the functions in the real atomic positions in order to obtain the real spins . in the case of an arbitrary choice of @xmath53
we should take the function values in the points shifted from the real atomic positions .
it is convenient to choose the parameter @xmath53 from the condition @xmath95 where @xmath96 means the direct product of two vectors .
notice that this @xmath97-tensor is proportional to the unit one due to the averaging over the symmetry elements of the point group @xmath28 ( see appendix , eq .
( [ eq : aijbij ] ) ) .
thus , for example , in the nearest neighbors approximation the condition gives @xmath98 , which is shown in ref .
@xcite to be necessary in order to obtain smooth spin functions . for @xmath99 using eqs .
( [ eq : sumtaub ] ) we find from eq .
( [ eq : x ] ) @xmath100 where the index `` exch '' means that @xmath101 is not a real coordinate , but some physical parameter expressed trough the exchange interaction constants . in untwisted spin structures ,
the canting is determined wholly by the dm interactions . in twisted states ,
an extra canting arises induced by a disagreement of phases of the helices connected with different magnetic sublattices .
the physical meaning of the `` exchange '' coordinates is that the spin shift to the new positions removes the canting appearing due to the spiralling .
the bond @xmath40 of the @xmath35th shell connects the function values @xmath102 and @xmath103 . using continuity of the functions
@xmath104 we can write @xmath105 where @xmath106 is the operator representing the taylor series expansion , and pass from the summation over the unit cells to the integration over the crystal volume ( in the chosen units , the lattice parameter @xmath107 and the unit cell volume is equal to 1 ) : @xmath108 @xmath109 \\ & - ( \mathbf{d}_{n , ij } \cdot \hat{\mathbf{s}}_i)(\mathbf{d}_{n , ij } \cdot { \cal b}_{n , ij } \hat{\mathbf{s}}_j)/(2j_n ) \ } - g \mu_b \mathbf{h } \cdot \sum_{i=1}^4 \hat{\mathbf{s}}_i . \end{array}\ ] ]
in the phenomenological theory , the magnetic moment density @xmath30 is used as an order parameter and considered as an experimentally observed physical quantity . nevertheless , when we try to express a smooth continuous function @xmath30 through a discrete distribution of spins of the magnetic atoms , an ambiguity arises from the fact that we can determine the weight function in different ways when averaging the spins on a local volume .
in this case the ambiguity results in that the smooth functions @xmath110 having specified values in a discrete set of points can be defined by infinite number of ways ( fig .
[ fig - xid](a ) )
. we will avoid the problem , supposing that the functions @xmath110 are defined in the most convenient way .
it would be natural to define the magnetization as @xmath111 @xmath112 in analogy with eq .
( [ eq : momentumb ] ) we introduce the canting tensor @xmath113 so that @xmath114 hereinafter the summation on repeated greek indices is implied .
four conditions @xmath115 , can be rewritten using the invariants @xmath116 @xmath117 in order to express the energy as a functional of the magnetic moment @xmath118 , the energy should be minimized by the canting tensor components using calculus of variations .
the variation of the energy is @xmath119 where the integrand now includes the lagrange terms : @xmath120 taking into account eq .
( [ eq : udef ] ) , @xmath121 m_\alpha \\ & -\tilde{j}_n [ \tau_{i\sigma } \tau_{j\rho } { \cal b}_{n , ij } + \tau_{j\sigma } \tau_{i\rho } { \cal b}_{n , ij}^{-1 } ] u_{\rho\alpha } \\ & -\varepsilon_{\alpha\beta\gamma } d_{n , ij\beta } [ \tau_{i\sigma } { \cal b}_{n , ij } - \tau_{j\sigma } { \cal b}_{n , ij}^{-1 } ] m_\gamma \\ & -\varepsilon_{\alpha\beta\gamma } d_{n , ij\beta } [ \tau_{i\sigma } \tau_{j\rho } { \cal b}_{n , ij } - \tau_{j\sigma } \tau_{i\rho } { \cal b}_{n , ij}^{-1 } ] u_{\rho\gamma } \\ & - ( 1/2j_n ) { d}_{n , ij\alpha } { d}_{n , ij\beta } [ \tau_{i\sigma } { \cal b}_{n , ij } + \tau_{j\sigma } { \cal b}_{n , ij}^{-1 } ] m_\beta \\ & - ( 1/2j_n ) { d}_{n , ij\alpha } { d}_{n , ij\beta } [ \tau_{i\sigma } \tau_{j\rho } { \cal b}_{n , ij } + \tau_{j\sigma } \tau_{i\rho } { \cal b}_{n , ij}^{-1 } ] u_{\rho\beta } \ } \\ & + 2 \lambda_0 u_{\sigma\alpha } + 2 \lambda_\sigma m_\alpha + 2 |\varepsilon_{\sigma\beta\gamma}| \lambda_\beta u_{\gamma\alpha } . \end{array}\ ] ] here @xmath122 and the evident equation was used , which corresponds to the integration in infinite volume , @xmath123 the minimum of the energy is determined from the condition @xmath124 with arbitrary functions @xmath125 , so the problem is reduced to the system of 9 equations @xmath126 with the extra conditions ( [ eq : invarianti0 ] ) , ( [ eq : invariantsi ] ) determining the functions @xmath127 , @xmath128 , @xmath129 , @xmath130 . the general solution of the problem is too difficult , if possible , but we can use the perturbation theory with a small parameter @xmath131 that is the ratio of typical absolute values of spin - orbit ( @xmath132 ) and exchange ( @xmath133 ) interactions .
we assume that this parameter also describes the typical order of magnitude of spatial derivatives and the spin components responsible for the canting : @xmath134 another quantity that can be connected with the small value of canting is @xmath135 . in a weak magnetic field , when a spiral structure still exists , @xmath136 .
but if the field is very strong ( @xmath137 @xcite ) , it induces a ferromagnetic alignment , and @xmath138 . in that case
we can take the maximal of two parameters , @xmath139 , as constitutive one . in the next section
, we will use the consecutive approximations in order to find a solution of the system ( [ eq : psi ] ) .
assuming @xmath140 , @xmath141 , the zeroth order equations on @xmath139 have the view @xmath142 and become trivial after the summation on bonds , see eq .
( [ eq : aij ] ) .
the first order equations on @xmath139 are @xmath143 where upper index @xmath144 means that the corresponding term is of the order of @xmath145 .
the first summand in curly brackets gives zero in accordance with eq .
( [ eq : x ] ) , two other can be calculated using eqs .
( [ eq : sumtautau]f ) and ( [ eq : td ] )
. then , @xmath146 the normalization conditions are @xmath147 @xmath148 where the first equation is of the second order on @xmath139 .
multiplication of eq .
( [ eq : approx1b ] ) by @xmath149 and summation on @xmath12 with use of eq .
( [ eq : norm1b ] ) give @xmath150 therefore @xmath151 the substitution of eq .
( [ eq : u1a ] ) into eq .
( [ eq : norm1a ] ) gives @xmath152 and , finally , @xmath153 where the sign of the right parts of eqs .
( [ eq : lambda01 ] ) and ( [ eq : u1b ] ) is chosen from the condition of the minimum of the canting energy .
the substitution of eq .
( [ eq : u1b ] ) into eq .
( [ eq : udef ] ) gives @xmath154 , \ ] ] with @xmath155 from eqs .
( [ eq : approx1b ] ) , ( [ eq : u1a ] ) and ( [ eq : u1b ] ) we conclude that responsible for the canting is the combination @xmath156 of the dm vectors components .
the contributions to the energy density from the magnetic moment @xmath118 and its derivatives are @xmath157 @xmath158 @xmath159 using eqs .
( [ eq : sumbb ] ) , ( [ eq : sumdb2 ] ) , and ( [ eq : dd ] ) with @xmath160 we obtain @xmath161 - \left ( \frac{2d_1 ^ 2}{j_1 } + \frac{2d_2 ^ 2}{j_2 } + \frac{2d_3 ^ 2}{j_3 } \right ) m^2 , \ ] ] with @xmath162 @xmath163 the first term in eq .
( [ eq : calem2 - 2 ] ) can be rewritten as @xmath164 where the second term gives a contribution to the surface energy only : @xmath165 far from the transition between paramagnetic and ferromagnetic states , the absolute value @xmath166 of the magnetic moment changes slowly , and this contribution to the energy can be neglected . however , near the phase transition , @xmath166 can undergo considerable changes , and in the crystals with a significant surface ( nanocrystals , thin films ) the term ( [ eq : esuf ] ) could play an important role . in particular , it could be important for the stabilization of the a - phase observed in thin films of fe@xmath167co@xmath167si and fege @xcite .
the contributions from the canting into the energy density are @xmath168 @xmath169 thus , we can finally rewrite the bulk energy density as a function of the magnetic moment @xmath118 accurate within the second order terms on @xmath139 : @xmath170 + ( \tilde{j}_1 + \tilde{j}_2 + \tilde{j}_3 ) ( 4 - 16 m^2 ) \\ & - 8 \sqrt{2 } |d_{1 + } + d_{2 + } + d_{3+}| m \sqrt{1-m^2 } - \left ( \frac{2d_1 ^ 2}{j_1 } + \frac{2d_2 ^ 2}{j_2 } + \frac{2d_3 ^ 2}{j_3 } \right ) m^2 - 4 g \mu_b \mathbf{h } \cdot \mathbf{m}. \end{array}\ ] ] the first two terms with derivatives are nothing but the deformation energy of the conventional phenomenological theory of the chiral magnetics .
however here the values @xmath8 and @xmath9 are expressed through the parameters of the microscopic theory accordingly to eqs .
( [ eq : calj ] ) and ( [ eq : cald ] ) .
the following term , on conditions that @xmath171 , gives @xmath172 , that is the energy of 36 ferromagnetic bonds in first three shells in the unit cell .
then , the term with @xmath135 is the contribution of the canting .
it is always negative , which means that the canting is an important and unavoidable peculiarity of the magnetic structures of the mnsi - type crystals . when minimizing the energy on @xmath166 , this term gives the contribution to the derivative proportional to @xmath173 , which becomes dominating , when @xmath174 , impeding @xmath166 to exceed 1 .
now we return to eq .
( [ eq : cald ] ) , which defines the dm parameter @xmath175 of the phenomenological theory . as it was mentioned above , from the physical point of view
, the dm vectors are almost perpendicular to the corresponding bonds and , consequently , the expected value of @xmath9 is small . in ref .
@xcite some speculative estimation has been performed with use of the well known keffer rule @xcite based on the moriya theory @xcite , which gives us following expression for the dm vector @xmath176 , \ ] ] where @xmath132 is unknown coefficient , and the vectors @xmath177 and @xmath178 are directed from the positions of 1st and 2nd magnetic ( mn ) atoms to that of an intermediate nonmagnetic ( si ) atom realizing the spin - orbit interaction .
it is evident that @xmath179 . for
the keffer rule one needs the constants @xmath132 and coordinates of different intermediate atoms .
surprisingly , a considerable result can be achieved with a more general rule , namely that the dm vectors are perpendicular to the bonds .
the condition can be written as @xmath180 and , comparing with eq .
( [ eq : cald ] ) , we easily find in this case that @xmath181 which gives us a new definition of the exchange coordinate as the manganese atom position inhibiting spiralling when combining with the keffer rule .
the combination @xmath182 is responsible for the canting in accordance with sec .
[ sec:1stapprox ] , and @xmath101 is a combination of the exchange interaction constants .
( [ eq : cald2 ] ) can be interpreted as an evidence of the canting direct participation in the magnetic structure spiralling . besides
, the sign of @xmath9 and consequently the magnetic chirality are determined both by the sign of the canting component @xmath183 of the dm vectors and that of the difference @xmath184 , depending on the exchange constants @xmath185 , @xmath186 , and @xmath187 .
therefore , the canting , initially considered as a supplementary microscopic peculiarity of the chiral magnetics @xcite , can be in fact an essential cause of the twisting power .
let us demonstrate by a simple , albeit not very realistic , example how the canting between different magnetic sublattices can result in an essential twist gain .
we consider a periodical 1d chain of spins with a local interaction between them , composed of unit cells containing two spins , say a and b ( fig . [ fig - chain ] ) .
all the spins can rotate in a plane , and the energy of the chain is a function of differences of the angles , @xmath188 here @xmath189 and @xmath190 are orientation angles of the spins a and b of the @xmath35th cell , laid off from an arbirary direction ; @xmath191 and @xmath192 are positive constants . the condition @xmath193 determines the chirality of the structure . when @xmath194 , eq .
( [ eq : echain ] ) describes a periodical magnetic structure with the angle @xmath195 of canting between the spin sublattices a and b ( fig . [ fig - chain](a ) ) .
we suppose that this structure is a result of the competition of two ordering interactions between neighboring spins , a ferromagnetic one and a twisting chiral one ; besides , there is a reason , which is not considered here , eliminating the twist in this state .
the question arises : what happens with the structure after introducing of an additional ferromagnetic ( @xmath196 ) interaction with the nearest neighbors from the complementary sublattice ?
the minimization of eq .
( [ eq : echain ] ) gives the solution in the helix form @xmath197 where @xmath198 is a tilt angle between spins in the unit cell , and @xmath199 is the twist angle per one period of the chain ( fig .
[ fig - chain](b ) ) . at the first sight it seems to be paradoxical that the introducing of an additional aligning interaction induces a twist , but a simple analysis of the problem shows that there is no contradiction here .
when @xmath200 , the angles @xmath12 and @xmath201 go to zero as expected . in order to understand the system behavior for the finite values of @xmath133 ,
consider the numerator of eq .
( [ eq : ddelta ] ) .
the factor @xmath202 reflects the degree of the internal chirality of the structure .
the product @xmath203 reminds of the lever torque , with @xmath195 playing the role of the lever arm and @xmath133 being analogous to the rotating force .
we can imagine that the `` aligning force '' @xmath133 , being applied to the initially tilted by the angle @xmath195 sublattices , results in structure distortions , which in their turn induce the twist due to the potential chirality of the structure ( @xmath193 ) .
notice that the change of the sign of the constant @xmath133 , corresponding to the transition from ferromagnetic ( @xmath196 ) to antiferromagnetic ( @xmath204 ) coupling , increases degree of the twist ( reduces the helix pitch ) and changes its handedness .
we expect that the similar effect takes place in the chiral magnetic mnsi .
indeed , there is an evident similarity of eqs .
( [ eq : cald2 ] ) and ( [ eq : ddelta ] ) . in this case
the role of a lever is played by the canting induced by the dm interactions , whereas the `` force '' is the ferro- or antiferromagnetic aligning interaction with neighboring atoms .
in analogy with the dzyaloshinskii moriya interaction , the exchange one is described differently in microscopical and phenomenological approaches .
although it is evident that exchange constants of both the theories should be connected to each other , the connection is found in the nearest neighbors approximation @xcite to be not so trivial as we could expect .
indeed , the expression ( [ eq : calj ] ) for @xmath8 in the approximation of three magnetic shells is strongly different from the simple proportionality in the former case .
however , the situation seems to be even more intricate , because @xmath8 is not sufficient to induce the ferromagnetic alignment , and another exchange parameter is also needed , namely @xmath205 where summation is taken over all the neighbors contributing to the exchange interaction .
then the condition of the validity of our approximation can be written as two inequalities , @xmath206 @xmath207 both equally important . here
@xmath208 is nothing but taken with an opposite sign energy of an individual spin interaction with its magnetic surroundings in the untwisted ( ferromagnetic ) state , and eq .
( [ eq : jsum0 ] ) guarantees stability of the state relative to the change of the spin sign .
( [ eq : jcal0 ] ) in its turn guarantees stability relative to the spin small rotations , providing smallness of the magnetic moment gradients
. therefore , only combined use of the conditions leads to a ferromagnetic ordering with a weak spiralling . neglecting the contribution of the dm interaction into symmetrical exchange ( i.e. using @xmath38 instead of @xmath209 ) , we can estimate @xmath210 and @xmath8 in the frame of the rkky theory @xcite , which is applicable to the itinerant magnetics .
indeed , in this model @xmath38 is a simple function of the distance between interacting atoms , @xmath211 @xmath212 here @xmath213 is the fermi wave number , @xmath214 is the dimensionless distance to the @xmath35th magnetic shell .
it follows from eqs .
( [ eq : jrkky ] ) , ( [ eq : f ] ) that atoms situated at approximately the same distances , e.g. @xmath215 and @xmath216 , make similar contributions to the exchange energy .
therefore , together with the 2nd and 3rd shells it is necessary to take into account the 4th shell corresponding to the atoms separated by lattice periods .
because the atoms of the 4th shell belong to the same magnetic sublattice , they do not affect the canting between different sublattices , i.e. they leave @xmath217 and @xmath101 unchanged and give simple additive contributions to @xmath210 and @xmath8 .
all the magnetic shells have 6 atoms , so @xmath218 in order to calculate the additive from the 4th shell to @xmath219 , we can write , for example , the interaction energy of two spins separated by the period @xmath220 of the lattice , @xmath221 if we take the sum over 6 bonds , multiply it by the number of magnetic atoms in the unit cell and take into account that all the bonds in that sum are taken twice , then the correction to the energy density can be written as @xmath222 it is obvious from the comparison with eq .
( [ eq : calem2 - 2 ] ) that the additive to @xmath8 from the 4th magnetic shell is @xmath223 . the parameters @xmath224 are oscillating functions of the argument @xmath225 , which are easy to calculate using real distances between magnetic atoms .
in mnsi for @xmath94 we find @xmath226 , @xmath227 , @xmath228 and @xmath229 .
[ fig - j](a ) shows the functions @xmath230@xmath231 plotted in the area @xmath232 , where the value @xmath233 for mnsi is situated ( @xmath234@xmath235 @xcite , @xmath236 ) .
the graphs show that eqs .
( [ eq : jsum0 ] ) , ( [ eq : jcal0 ] ) can be satisfied together in the areas of negative values of @xmath230 .
[ fig - j](b ) represents dependences of @xmath210 and @xmath8 on @xmath225 , calculated for the same area .
both @xmath210 and @xmath8 are oscillating alternating - sign functions ; the jumps of @xmath8 correspond to the zeros of @xmath237 .
the nearest to the value @xmath238 `` plateau '' with positive @xmath210 and @xmath8 is in the area @xmath239 .
notice that the behavior of @xmath8 at some @xmath240 could seem paradoxical .
for example , the divergence of @xmath241 would mean a strong suppression of long - wavelength fluctuations .
however , the paradox can be resolved by taking into account the behavior of @xmath210 . indeed , if @xmath242 , then @xmath243 , where @xmath244 is nothing but the coupling constant of the spins belonging to the same sublattice .
this means that , when @xmath241 , the connection between sublattices gets broken .
thus , when the spins of three sublattices are aligned in the same direction , the spins of the fourth one can have an arbitrary direction , even if the condition @xmath245 guarantees the ferromagnetic order within the sublattice .
the foregoing means that in addition to eqs .
( [ eq : jsum0 ] ) , ( [ eq : jcal0 ] ) , we should introduce another inequality , @xmath246 determining the ferromagnetic connection of the four magnetic sublattices . here the sum is taken over the bonds connecting atoms belonging to different sublattices . in the approximation of four magnetic shells , @xmath247 , therefore the zeros of @xmath248 coincide with the jumps of @xmath8 .
the rkky model allows as well to estimate the exchange coordinate @xmath101 .
because , according to eq .
( [ eq : cald2 ] ) , the degree of the twist is determined by the difference @xmath249 , it is useful to have an idea about how much this difference could be . in the nearest neighbors approximation @xmath250 , which is close to the real value @xmath251 for mnsi .
however , when taking into account the contributions from the 2nd and 3rd shells , @xmath101 can have arbitrary large positive and negative values near the zeros of @xmath237 ( fig .
[ fig - xexch ] ) .
close to the minima of @xmath230 @xmath252 and @xmath253 are small in comparison with @xmath254 and , therefore , @xmath255 , see inset in fig .
[ fig - xexch ] .
notice that the zeros of @xmath237 do not result in a divergence of the wave number @xmath256 , because @xmath8 in the denominator and @xmath101 in the numerator increase simultaneously . as it is seen from fig .
[ fig - xexch ] and the inset in it , the difference @xmath257 can change its sign depending on @xmath213 .
it gives a possibility to control the magnetic structure chirality by varying the concentration of different elements in the crystal .
( [ eq : jsum0 ] ) , ( [ eq : jcal0 ] ) are evident preconditions of the experimentally observed ferromagnetic order in mnsi .
nevertheless , we can not preclude that one of the constants @xmath210 , @xmath8 , or both these parameters can have negative values .
for example , the condition @xmath258 does not surely result in an antiferromagnetic order .
the strong frustrations intrinsic for the system , e.g. the triangles of bonds , and nonsymmetric dm interactions can induce a small magnetic moment and lead to a ferri- or a weak ferromagnetic order .
it could explain the weak magnetic moment observed in mnsi ( @xmath259 ) .
when @xmath260 , the contributions to the energy density with higher spatial derivatives should be taken into account , which can stabilize the helix pitch .
in sec . [ sec:1stapprox ] , we obtain eq .
( [ eq : shat ] ) , which says that in the first approximation the canting can be described as spin rotations by the same small angle around corresponding 3-fold axes .
a similar expression for the ferromagnetic state caused by a strong magnetic field was found in ref .
@xcite , where an approach had been suggested to measure the canting using neutron or x - ray magnetic diffraction . in ref .
@xcite , the angle of spin tilt was proportional to @xmath261 . in order to obtain that result
we can find the coefficient @xmath262 for untwisted state .
the minimization of eq .
( [ eq : cale ] ) on magnetic moment modulus @xmath166 , assuming that @xmath263 , gives @xmath264 notice that @xmath265 , when @xmath266 .
the corresponding expressions for the structure factors of purely magnetic reflections @xmath267 can be easily found from ref .
@xcite .
notice that there is another contribution to the reflections , induced by the anisotropy of magnetic susceptibility tensor @xcite .
the contributions can be distinguished , because ( i ) the tilts have different directions , ( ii ) the canting effect does not depend on the magnetic field modulus , whereas the tilts induced by the susceptibility tensor anisotropy are proportional to @xmath268 .
the fact that both the twist and canting are determined by the same dm vectors components gives an additional possibility of numerical verification of the theory .
indeed , excluding @xmath183 from eqs .
( [ eq : kappa2 ] ) and ( [ eq : cald2 ] ) , we can connect the canting angle in unwound state and the wave number @xmath269 of magnetic helices : @xmath270 excepting observable physical values , eq . ( [ eq : kappa3 ] ) contains only exchange interaction constants , which are easier to calculate using _
ab initio _ methods than the dm vectors .
an essential difference of the microscopic description of the magnetic properties of the mnsi - type crystals from the phenomenological one is the usage of the pseudovector @xmath271 instead of pseudoscalar @xmath9 when describing the dzyaloshinskii
moriya interaction .
the presence of the extra parameters ( pseudovector components ) results in the existence of a local canting , the feature not studied yet in magnetic twisted structures .
an important problem solved in the present work is how to distinguish the components of @xmath271 responsible for the twist and the canting .
hopkinson and kee in ref .
@xcite showed numerically ( in the nearest neighbors approximation ) that responsible for the twist were the components of the @xmath271-vectors lying along the bonds , whereas for the canting did the @xmath271 components directed perpendicular to the bonds and lying in the planes of bond triangles in the trillium lattice . in ref .
@xcite we specified the result and showed that real components of @xmath271 inducing the twist and the canting lay along crystallographic directions closed to those found in ref .
@xcite . nevertheless , the problem remained that in accordance with the quantum mechanical description the dm vectors should be perpendicular or almost perpendicular to the bonds . in other words , the @xmath271 components responsible for the twist accordingly to refs .
@xcite could be diminutive . in order to solve the problem we take into account the contribution of non - nearest neighbors in the magnetic interaction .
surprisingly , it appears that in the case , when all the dm vectors are perpendicular to the bonds , the spiralling is determined by the same @xmath271 components as the canting .
it leads to the conclusion that the canting , initially being considered as an additional microscopic effect in relation to the global twist , can in fact serve as a cause of the abnormal twist experimentally found in the mnsi - type crystals , particularly in mnge .
it is also important that the contribution of non - nearest magnetic neighbors should be taken into account . in the simplest phenomenological theory describing twisted magnetic structures
, it is supposed that @xmath272 , which is roughly true at low temperatures , far below the phase transition from the paramagnetic state .
however , this condition makes energetically unfavorable such structures as the skyrmions and their lattices , associated with the a - phase observed close by the transition point . in order to overcome this problem as well as to describe the critical phenomena , two additional terms , @xmath273 and @xmath274 , limiting the value of magnetic moment
are included into the free energy @xcite .
the presence of the terms decreases @xmath275 in the regions with a large density of the magnetic energy , thereby decreasing the energy of the whole structure .
thus , in ref .
@xcite the terms @xmath273 and @xmath274 are used in order to calculate the energy of skyrmions .
it is found that @xmath275 decreases considerably nearby the core of the skyrmion , where the magnetic moment @xmath30 has the opposite direction to the external magnetic field , and has a maximum at some distance from the core , where the energy gain from the double twist is maximal .
however , in the latter case the magnetic moment exceeds its saturation value @xmath276 , which is not acceptable for physical reasons . in the present work
we show that there should be a contribution from the canting @xmath277 into the energy density , which can play the same role as @xmath273 and @xmath274 , but does not allow the magnetic moment modulus to exceed the saturation threshold . besides the canting , the thermal fluctuations of spins also contribute to the reduction of the magnetic moment @xmath275 .
nearby the transition point the amplitude of the fluctuations can be comparable with the canting . moreover , the less is the effective local field @xmath278 , acting on the individual spin @xmath41 , the more are fluctuations and cantings .
therefore , the reinforcement both of the thermal fluctuations and the canting have the same cause , so they should give a similar effect .
if the canting would be observed in mnsi , the direct confirmation of non - nearest neighbors effect could be possible using eq .
( [ eq : kappa2 ] ) , connecting the propagation number of the magnetic helices with the magnitude of the residual canting in the unwound state in a strong magnetic field .
the equation involves only the exchange constants @xmath38 and does not depend on the dm vectors . the possibility of an experimental proof of the theory should stimulate _ ab intio _ calculations of the interaction constants .
some semi - quantitative estimations are made in the present work with use of the rkky model .
however problems still remain .
for example , the rkky parameter @xmath238 for mnsi corresponds to the area , where @xmath258 and @xmath279 .
more realistic calculations would give a tip about the direction of the further search .
it has been found in ref .
@xcite that the propagation number @xmath280 of the magnetic helix in the alloy fe@xmath1co@xmath281si is strongly dependent on the concentration @xmath53 of the cobalt atoms .
thus , when @xmath53 changes from 0.05 to 0.15 , the helix period decreases abruptly in several times .
it can be explained , in particular , by the strong dependence of the fermi wave number @xmath213 on the cobalt impurity concentration , which also effects on the exchange parameters @xmath210 , @xmath8 and @xmath101 .
this phenomenon could be also responsible for the recently observed @xcite change of the sign of the magnetic chirality in mn@xmath1fe@xmath2ge alloys .
this is drastically different from the usually supposed change of the sign of the dm interaction . indeed , eq .
( [ eq : cald2 ] ) shows that the chirality can change even if the microscopic dm interaction , defined by the vectors @xmath271 , remains constant . in this case
the sign change is due to the interplay between the exchange parameters @xmath254 , @xmath252 , and @xmath253 of three magnetic shells .
thus a potential possibility arises to control the sign of the magnetic chirality by varying the concentration of the different components and therefore affecting the fermi wave number @xmath213 .
notice that the possibility of such effect becomes evident only when the interactions with non - nearest neighbors are taken into account .
another interesting fact , not being yet explained within the framework of microscopic theories , is the helix ordering along some special directions , e.g. @xmath282 or @xmath283 . usually this ordering is associated with a weak anisotropic exchange @xcite , but in fact the ordinary dm interaction also can result in the appearing of cubic anisotropic terms in the energy of spiral orientation in the lattice with a cubic space group .
these contributions of the order @xmath284 will determine the critical magnetic field @xmath285 , at which the helix comes off from its preferable zero - field direction .
this is in a good agreement with the observed ratio of the first and second critical fields @xmath286 , because it follows from our estimations that @xmath287 . indeed , e.g. the period of the magnetic helix in mnsi makes about 40 unit cell parameters , which gives the value @xmath288 for the propagation number modulus @xmath289 . on the other part , it follows from the phenomenological description that @xmath290 , or @xmath291 .
this gives us the estimation @xmath292 , which is in a good agreement with the experimental data . in our previous work
@xcite we proposed a coarse grain approximation , which allowed us to calculate only the contributions to the energy proportional to @xmath20 . in the present work
a new approach has been developed permitting more precise calculations of the energy .
in particular , the terms of the order of @xmath284 would give us a contribution of the dzyaloshinskii
moriya interaction into the energy of the cubic anisotropy responsible for ( i ) the ordering of the spiral axes along selected crystallographic directions , e.g. @xmath282 in mnsi , in the absence of external magnetic field ; ( ii ) the orientation of the triangle skyrmion lattice in the a - phase observed in these crystals @xcite .
we are grateful to m. v. gorkunov , s. v. demishev , s. v. maleyev , s. v. grigoriev , v. a. dyadkin and s. s. doudoukine for useful discussions and encouragement .
this work was supported by two projects of the presidium of the russian academy of sciences : `` matter at high energy densities ; substance under high static compression '' and `` diffraction of synchrotron radiation in multiferroics and chiral magnetics '' .
[ [ appendixa ] ] the vectors @xmath82 , @xmath293 , @xmath294 , @xmath295 , @xmath296 $ ] , @xmath297 , @xmath39 ( @xmath99 ) change with use of the same symmetry transformations of the point group @xmath28 , when the index @xmath40 passes trough 12 possible values @xmath298 .
therefore we can use the formulae @xmath299
@xmath300 where @xmath301 and @xmath75 are the vectors from the above - listed set .
thereby the following summations can be easily made : [ eq : sumtautau ] @xmath302_{\beta } = 0 , \\
\sum_{\{ij\ } } ( \boldsymbol{\tau}_i - \boldsymbol{\tau}_j)_{\alpha } ( \boldsymbol{\tau}_i - \boldsymbol{\tau}_j)_{\beta } = 32 \delta_{\alpha\beta } , \\
\sum_{\{ij\ } } ( \boldsymbol{\tau}_i + \boldsymbol{\tau}_j)_{\alpha } ( \boldsymbol{\tau}_i + \boldsymbol{\tau}_j)_{\beta } = 16 \delta_{\alpha\beta } , \\
\sum_{\{ij\ } } [ \boldsymbol{\tau}_i \times \boldsymbol{\tau}_j]_{\alpha } [ \boldsymbol{\tau}_i \times \boldsymbol{\tau}_j]_{\beta } = 32 \delta_{\alpha\beta } , \\
\sum_{\{ij\ } } ( \tau_{i\alpha } \tau_{j\beta } + \tau_{j\alpha } \tau_{i\beta } ) = -8\delta_{\alpha\beta } .\end{aligned}\ ] ] using eqs .
( [ eq : sumtautau]a)-([eq : sumtautau]e ) we can calculate the sums containing products of the vectors @xmath297 , @xmath39 ( @xmath99 ) : [ eq : sumtaub ] @xmath303 @xmath304 @xmath305 [ eq : sumbb ] @xmath306 [ eq : sumdb2 ] @xmath307 99 h. j. williams , j. h. wernick , r. c. sherwood , and g. k. wertheim , j. appl . phys . * 37 * , 1256 ( 1966 ) .
d. shinoda and s. asanabe , j. phys .
japan * 21 * , 555 ( 1966 ) .
s. m. stishov and a. e. petrova , physics - uspekhi * 54 * , 1117 ( 2011 ) .
y. ishikawa , k. tajima , d. bloch , and m. roth , solid state commun . * 19 * , 525 ( 1976 ) .
k. motoya , h. yasuoka , y. nakarnura , and j. h. wernick , solid state commun .
* 19 * , 529 ( 1976 ) .
u. k. rler , a. v. bogdanov , and c. pfleiderer , nature ( london ) * 442 * , 797 ( 2006 ) .
s. v. grigoriev , s. v. maleyev , a. i. okorokov , yu .
o. chetverikov , p. bni , r. georgii , d. lamago , h. eckerlebe , and k. pranzas , phys .
b * 74 * , 214414 ( 2006 ) .
w. mnzer , a. neubauer , t. adams , s. mhlbauer , c. franz , f. jonietz , r. georgii , p. bni , b. pedersen , m. schmidt , a. rosch , and c. pfleiderer , phys .
b * 81 * , 041203 ( 2010 ) .
t. adams , s. mhlbauer , c. pfleiderer , f. jonietz , a. bauer , a. neubauer , r. georgii , p. bni , u. keiderling , k. everschor , m. garst , and a. rosch , phys .
lett . * 107 * , 217206 ( 2011 ) .
s. d. tewari , d. belitz , and t. r. kirkpatrick , phys .
lett . * 96 * , 047207 ( 2006 ) .
b. binz , a. vishwanath , and v. aji , phys .
* 96 * , 207202 ( 2006 ) .
a. hamann , d. lamago , th .
wolf , h. v. lhneysen , and d. reznik , phys .
* 107 * , 037207 ( 2011 ) .
j. m. hopkinson and h .- y .
kee , phys . rev .
b * 74 * , 224441 ( 2006 ) . v. a. chizhikov and v. e. dmitrienko , phys . rev .
b * 85 * , 014421 ( 2012 ) .
i. dzyaloshinsky , phys .
solids * 4 * , 241 ( 1958 ) .
p. bak and m. h. jensen , j. phys .
c : solid state phys . *
13 * , l881 ( 1980 ) .
o. nakanishi , a. yanase , a. hasegawat , and m. kataoka , solid state commun . * 35 * , 995 ( 1980 ) .
t. moriya , phys .
rev . * 120 * , 91 ( 1960 ) .
j. m. hopkinson and h .- y .
kee , phys .
b * 79 * , 014421 ( 2009 ) . v. e. dmitrienko and v. a. chizhikov , phys . rev .
* 108 * , 187203 ( 2012 ) .
s. seki , x. z. yu , s. ishiwata , and y. tokura , science * 336 * , 198 ( 2012 ) .
t. adams , a. chacon , m. wagner , a. bauer , g. brandl , b. pedersen , h. berger , p. lemmens , and c. pfleiderer , phys .
lett . * 108 * , 237204 ( 2012 ) .
j. h. yang , z. l. li , x. z. lu , v .- h .
whangbo , su - huai wei , x. g. gong , and h. j. xiang , phys .
lett . * 109 * , 107203 ( 2012 ) .
a. k. zvezdin and a. p. pyatakov , eur .
* 99 * , 57003 ( 2012 ) .
s. a. pikin and i. s. lyubutin , phys .
b * 86 * , 064414 ( 2012 ) .
v. v. mazurenko and v. i. anisimov , phys .
b * 71 * , 184434 ( 2005 )
. e. j. samuelsen and g. shirane , phys .
status solidi * 42 * , 241 ( 1970 ) .
m. kenzelmann , a. b. harris , s. jonas , c. broholm , j. schefer , s. b. kim , c. l. zhang , s .- w .
cheong , o. p. vajk , and j. w. lynn , phys .
lett . * 95 * , 087206 ( 2005 ) .
cheong and m. mostovoy , nat .
* 6 * , 13 ( 2007 ) .
j. m. hopkinson and h .- y .
kee , phys .
b * 75 * , 064430 ( 2007 )
. p. j. brown , j. b. forsyth , and g. h. lander , j. appl . phys . * 39 * , 1331 ( 1968 ) .
a. s. moskvin and i. g. bostrem , fiz .
tela ( leningrad ) * 19 * , 2616 ( 1977 ) [ sov .
solid state * 19 * , 1532 ( 1977 ) ] .
l. shekhtman , a. aharony , and o. entin - wohlman , phys .
b * 47 * , 174 ( 1993 ) . i. a. sergienko and e. dagotto , phys . rev .
b * 73 * , 094434 ( 2006 ) .
m. i. katsnelson , y. o. kvashnin , v. v. mazurenko , and a. i. lichtenstein , phys .
b * 82 * , 100403 ( 2010 ) .
l. shekhtman , o. entin - wohlman , and a. aharony , phys .
69 * , 836 ( 1992 ) .
t. yildirim , a. b. harris , a. aharony , and o. entin - wohlman , phys .
b * 52 * , 10239 ( 1995 ) .
t. hahn ( ed . ) , _ international tables for crystallography , vol .
a : space - group symmetry _ ( kluwer academic , dordrecht , the netherlands , 1989 ) .
v. dmitriev , d. chernyshov , s. grigoriev , and v. dyadkin , j. phys . : condens
. matter , * 24 * , 366005 ( 2012 ) .
z. yu , y. onose , n. kanazawa , j. h. park , j. h. han , y. matsui , n. nagaosa , and y. tokura , nature * 465 * , 901 ( 2010 ) . x. z. yu , n. kanazawa , y. onose , k. kimoto , w. z. zhang , s. ishiwata , y. matsui , and y. tokura , nat . mater .
* 10 * , 106 ( 2011 ) .
f. keffer , phys . rev . * 126 * , 896 ( 1962 ) .
m. a. ruderman and c. kittel , phys .
* 96 * , 99 ( 1954 ) .
t. kasuya , prog .
* 16 * , 45 ( 1956 ) .
k. yosida , phys . rev . *
106 * , 893 ( 1957 ) .
j. h. van vleck , rev .
phys . * 34 * , 681 ( 1962 ) .
t. r. kirkpatrick and d. belitz , phys .
b * 80 * , 220401(r ) ( 2009 ) . a. gukasov and p. j. brown , j. phys . : condens .
matter * 14 * , 8831 ( 2002 ) .
h. cao , a. gukasov , i. mirebeau , p. bonville , c. decorse , and g. dhalenne , phys .
lett . * 103 * , 056402 ( 2009 ) .
u. k. rler , a. a. leonov , and a. v. bogdanov , j. phys .
: conf . ser .
* 303 * , 012105 ( 2011 ) .
s. v. grigoriev , d. chernyshov , v. a. dyadkin , v. dmitriev , s. v. maleyev , e. v. moskvin , d. menzel , j. schoenes , and h. eckerlebe , phys .
lett . * 102 * , 037204 ( 2009 ) .
s. v. grigoriev , n. m. potapova , s .- a .
siegfried , v. a. dyadkin , e. v. moskvin , v. dmitriev , d. menzel , c. d. dewhurst , d. chernyshov , r. a. sadykov , l. n. fomicheva , and a. v. tsvyashchenko , in press .
( color online ) . ( a ) the 1d twisted magnetic structure ; the spins rotate in the plane perpendicular to the propagation vector .
( b ) the canting in the absence of a twist ; the lattice is divided into two sublattices ( blue and sand - coloured arrows ) with the spins tilted by a constant angle .
( c ) the canting and the twist ; the canting of the sublattices leads to the combination of two helices with the same propagation vector and a constant phase shift.,width=264 ] ( color online ) .
nearest magnetic neighborhood of a manganese atom in mnsi .
the central atom ( magenta ) belongs to the first manganese sublattice , its neighbors are in 2nd ( green ) , 3rd ( orange ) and 4th ( blue ) ones .
the shortest bonds are given by solid lines , the bonds with the atoms from the 2nd and 3rd magnetic shells are shown by dashes and dots , correspondingly .
( a ) the 3-fold axis @xmath308 $ ] is vertical .
the horizontal lines designate equidistant atomic planes containing the manganese atoms of 2nd , 3rd and 4th sublattices .
( b ) the projection on the plane perpendicular to the axis @xmath308 $ ] . two atomic triangles from ( a ) ,
the upper and the lowest , belonging to different shells are situated just one above the other.,width=264 ] ( color online ) .
( a ) the ambiguity of the transition from discrete to continuous model : two different smooth functions having the equal values in a discrete set of points .
( b ) the gain from the arbitrary choice of the parameter @xmath53 .
the initial curve ( orange ) is plotted with the use of the function values given in the discrete set of points with the coordinates @xmath53 and @xmath309 within the cells of a 1d crystal . when the points are shifted to the center of the cells without a change of function values , the view of the curve changes ( blue).,width=264 ] ( color online ) .
periodical 1d chain of magnetic atoms with two spins within a unit cell .
the spins can rotate in a plane , and their directions are fully determined by the only variable @xmath310 .
the spin interaction is described by the classic hamiltonian ( [ eq : echain ] ) .
( a ) when @xmath194 , the chirality is hidden ; the lattice is divided onto two sublattices with the ferromagnetic ordering of spins and the canting angle @xmath195 between them .
( b ) when an additional ferromagnetic interaction with the atoms of the alternate kind is included ( @xmath311 ) , the chirality becomes apparent in the magnetic helix with the angle @xmath201 between neighboring unit cells ; besides , the angle @xmath12 between the atoms @xmath3 and @xmath312 within a cell remains the same over the chain.,width=264 ] ( color online ) . ( a )
the @xmath313 dependences on the fermi wave number for four magnetic shells .
the distances @xmath314@xmath315 are chosen as for mnsi .
the minima of @xmath230 give the areas with @xmath316 .
( b ) the @xmath210 ( solid line ) and @xmath8 ( dash line ) dependences on the fermi wave number .
the plots show strong oscillations .
the jumps of @xmath8 correspond to the zeros of @xmath317 .
the areas with @xmath316 determine weakly twisted ferromagnetic states .
the most close to the known for mnsi value @xmath238 plateau is in the area @xmath239.,width=264 ] ( color online ) .
exchange coordinate @xmath101 calculated in the rkky model .
the jumps of the function correspond to the zeros of @xmath317 .
the dot line identifies the real value @xmath318 for mnsi .
the inset shows the @xmath101 dependence in the corresponding to positive @xmath210 and @xmath8 area @xmath239 , where @xmath319 .
the chirality of the magnetic structure change its sign , when @xmath320.,width=264 ] | the transition from the microscopic heisenberg model to the macroscopic elastic theory is carried out for the chiral magnetics of mnsi - type with the @xmath0 crystal structure .
both exchange and dzyaloshinskii moriya ( dm ) interactions are taken into account for the first , second , and third magnetic neighbors .
the particular components of the dm vectors of bonds are found , which are responsible for ( i ) the global magnetic twist and ( ii ) the canting between four different spin sublattices .
a possible mechanism for effective reinforcement of the global magnetic twist is suggested : it is demonstrated that the components of the dm vectors normal to corresponding interatomic bonds become very important for the twisting power .
the ruderman
kittel kasuya
yosida ( rkky ) theory is used for model calculation of exchange parameters .
it is found that just the interplay between the exchange parameters of several magnetic shells rather than the signs of dm vectors can be responsible for the concentration - induced reverse of the magnetic chirality recently observed in the mn@xmath1fe@xmath2ge crystals . |
density fluctuations on au - size scales in the interstellar plasma may produce a host of observable effects including refractive intensity scintillations and image wander .
the magnitude of these effects depends crucially upon the spectrum of the interstellar density fluctuations .
if the spectrum is a power law with a spectral index less than 4 , these refractive effects will generally be masked by diffractive effects from even smaller scale density fluctuations @xcite .
conversely , if the spectrum is a power law with a spectral index greater than 4 , refractive effects will dominate .
for many lines of sight the density spectral index appears to be close to , but less than , 4 @xcite , with important exceptions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
complicating the effort to measure refractive effects is their generally long time scale .
the relevant time scale is given by the time it takes for an au - sized density fluctuation to drift past the observer .
for observations of pulsars , whose velocities can exceed 100 , the time scale is days . for observations of extragalactic sources ,
the relevant velocity is a combination of the earth s velocity and random interstellar motions and is more likely to be roughly 25 , meaning that the refractive time scale is months to years .
the source is an ideal source for probing refractive effects .
it lies on the outskirts of the cygnus superbubble @xcite , a region known to exhibit enhanced interstellar scattering @xcite .
vlbi observations reveal its milliarcsecond structure to be dominated by two components separated by approximately 60 mas , thus a search for refractive image wander requires only _ relative _ position measurements .
* hereafter @xcite ) obtained 0.61 ghz vlbi observations of .
combined with previous measurements in the literature , they showed that scattering is moderately strong along this line of sight .
they were also able to constrain the amount of differential refractive image wander between the two components .
their constraint , however , is based on measurements of the separation of the components at two epochs separated by four years .
if the refractive time scale is longer than four years , their constraint is weakened considerably . assuming that typical interstellar velocities along this line of sight are 50100
, @xcite argued that the refractive time scale is comparable to or less than four years .
obviously , if the typical interstellar velocity is more like 25 , the refractive time scale is much longer .
we have acquired a third epoch of 0.61 ghz vlbi observations of , specifically to confront the conclusions of @xcite with another epoch of observations .
this third epoch is 12 to 16 yrs after the first two epochs and should allow us to place much better constraints on any refractive wander .
we have also acquired simultaneous or quasi - simultaneous observations at a range of other wavelengths in order to confirm previous measurements of the strength of interstellar scattering along this line of sight . in [ sec : observe ] we describe our observations and data analysis , in
[ sec : scatter ] we discuss the implications of our observations for both diffractive and refractive interstellar scattering along this line of sight , and in [ sec : intrinsic ] we discuss limits our observations place on the intrinsic structure and kinematics of the source .
we present our conclusions in [ sec : conclude ] .
we observed with the vlba for a total of 10 hours on 2000 july 25 .
relevant observation parameters are summarized in table [ tab : log ] .
0.33 & 12 & r , l & 2.88 & 61 @xmath1 58 @ 87 & 3.2 + 0.61 & 4 & r , l & 2.88 & 27 @xmath1 19 @ 14 & 5 + 1.67 & 32 & r , l & 2.93 & 7.3 @xmath1 4.7 @ @xmath2 & 0.78 + 2.27 & 16 & r , l & 2.92 & 5.8 @xmath1 3.6 @ @xmath3 & 1.1 + 8.42 & 16 & r , l & 2.92 & 1.9 @xmath1 1.4 @ @xmath4 & 0.61 + the observation frequencies were 0.33 , 0.61 , 1.67 , 2.3 , and 8.4 ghz ; both right and left circular polarization were recorded at all frequencies .
the 2.3 and 8.4 ghz observations were obtained simultaneously in `` s / x mode . ''
simultaneous 0.33 and 0.61 ghz observations also were obtained by setting two of the eight baseband converters ( bbcs ) at each vlba antenna to record the two senses of circular polarization at 0.61 ghz . because of severe radio frequency interference around 0.61 ghz , filters restricted the observable bandwidth for each bbc to 4 mhz .
the remaining bbcs recorded 0.33 ghz .
formatter problems at the kitt peak antenna meant that only nine of the ten vlba antennas were correlated .
the correlated visibilities were processed in the standard fashion within the nrao astronomical image processing system ( aips ) .
fringe - fitting intervals ranged from 10 min . at 1.67 ghz to 3 min.at 0.33 and 8.4 ghz .
at 0.33 and 0.61 ghz , substantial editing of the calibrated visibilities was required to remove radio frequency interference . in general the calibrated visibilities
were then exported to the caltech ` difmap ` program for hybrid mapping . to ensure a robust estimate for the separation of the components ( [ sec : ref ] ) , the 0.61 ghz observations were imaged using both aips and ` difmap ` ; the 0.33 ghz observations were imaged using only aips .
figure [ fig : images ] shows the images of at the various frequencies .
the observed rms noise level in these images ranges from 0.61 at 8.42 ghz to 5 at 0.61 ghz and are summarized in table [ tab : log ] .
the rms noise levels in our images are roughly within a factor of 23 of the expected value . using the images shown in figure [ fig : images ] as a guide , we fit the self - calibrated visibility data with multiple gaussian component models .
the number of gaussian components fitted ranged from 1 at 0.33 ghz to 7 at 8.42 ghz .
these models are summarized in table [ tab : models ] .
below we shall make use of the fitted angular diameters . at the higher frequencies ,
we take the uncertainty in a fitted angular diameter to be one half of the synthesized beam ; at the lower frequencies , where a one- or two - component fit suffices , we take the uncertainties to be the square roots of the diagonal elements of the covariance matrix obtained from a least - squares fit to the visibility data ( provided by the ` uvfit ` fitting routine within aips ) ; these values are of order 0.5 mas .
lcccccccc 0.33 & 2.17 & 0.0 & & 106.2 & 0.76 & @xmath5 & e + 0.61 & 2.48 & 0.0 & & 24.1 & 0.82 & 24 & e + & 0.39 & 56.1 & @xmath6 & 26.2 & 0.69 & 8 & w + & 0.11 & 32.6 & 8 & 47.5 & 0.00 & 67 & e + 1.67 & 1.75 & 59.3 & @xmath7 & 3.5 & 0.90 & @xmath8 & w + & 0.79 & 5.5 & @xmath6 & 4.2 & 0.74 & @xmath9 & e + & 0.99 & 1.3 & @xmath10 & 4.0 & 0.86 & 33 & e + & 1.06 & 3.4 & 42 & 4.9 & 0.69 & @xmath11 & e + & 0.16 & 8.9 & @xmath12 & 11.8 & 0.32 & @xmath13 & e + 2.27 & 1.97 & 0.0 & @xmath14 & 2.9 & 0.71 & @xmath15 & w + & 0.92 & 58.5 & 101 & 3.7 & 0.81 & 17 & e + & 0.58 & 53.8 & 100 & 3.3 & 0.59 & @xmath16 & e + & 0.83 & 61.5 & 97 & 4.2 & 0.54 & @xmath17 & e + & 0.12 & 52.9 & 103 & 7.5 & 0.39 & @xmath18 & e + 8.42 & 0.35 & 18.4 & @xmath19 & 0.8 & 0.54 & 84 & w + & 0.39 & 20.1 & @xmath20 & 1.0 & 0.51 & @xmath21 & w + & 0.19 & 34.9 & 103 & 2.1 & 0.64 & @xmath22 & e + & 0.19 & 40.3 & 105 & 2.1 & 0.54 & 60 & e + & 0.19 & 43.2 & 99 & 3.5 & 0.51 & @xmath23 & e + & 0.22 & 18.8 & @xmath23 & 3.1 & 0.24 & @xmath24 & w + & 0.16 & 39.0 & 102.4 & 6.3 & 0.28 & 40 & e +
we now revisit the scattering properties of the line of sight to .
we consider both diffractive effects , namely angular broadening , and refractive effects , namely differential image wander and refractive intensity fluctuations .
we shall make the standard assumption throughout that the density fluctuations responsible for interstellar scattering can be parameterized by an exponentially - truncated power law @xmath25 here @xmath26 is the largest wavenumber on which density fluctuations occur , which is related to the smallest length scales for density fluctuations @xmath27 or `` inner scale '' as @xmath28 .
the smallest wavenumber on which density fluctuations occur or the `` outer scale '' is @xmath29 . the constant @xmath30 is a measure of the level of density fluctuations and is related to the rms density as @xmath31\,{\ensuremath{c_n^2}}\ , l_0^{3-\alpha}$ ] @xcite .
the power law spectral index is assumed to provide clues about the process(es ) that generate or maintain the density fluctuations .
our observations allow us to measure @xmath32 , constrain @xmath30 , and indirectly place limits on @xmath27 .
an infinitely - distant point source viewed through a medium filled with density fluctuations having a spectrum described by equation ( [ eqn : spectrum ] ) and a spectral index @xmath33 will be broadened to an angular diameter @xcite @xmath34 here @xmath35 is the observing frequency measured in ghz , and the scattering measure sm is the line of sight integral of @xmath30 , @xmath36 the quantity @xmath37 is the cross - over angle and describes the relative importance of the inner scale @xmath27 .
it is given by @xmath38 .
the constants ( 128 mas and 132 mas ) given in equation ( [ eqn : thd ] ) are those appropriate for infinitely distant sources and differ slightly from those given by @xcite who considered sources embedded in the medium .
these formulae specify the diameter as the full width at half maximum ( fwhm ) of the shape of the image and that shall be the convention we adopt in reporting the measured diameters .
the scattering diameter @xmath39 defines the diffractive scale @xmath40 , @xmath41 ( at observing wavelength @xmath42 ) .
if the typical interferometer baseline @xmath43 is sufficient to resolve @xmath39 ( i.e. , @xmath44 ) and @xmath45 , we expect @xmath46 and @xmath47 .
conversely if @xmath48 , then we expect @xmath49 and @xmath50 . therefore fitting the observed diameter as a function of frequency can constrain both @xmath27 and @xmath32 .
there are alternate causes for a @xmath47 dependence .
for instance , if the spectrum of density fluctuations is not a power law but a gaussian centered on a particular length scale , that can produce a @xmath47 dependence .
similarly , if the density spectrum is dominated just slightly by au - size scales such that @xmath51 , that could also produce an apparent @xmath47 dependence within the measurement uncertainties . as we shall show below , @xmath52 so that we could interpret a @xmath47 dependence as resulting from a `` large '' inner scale . figure [ fig : diameter ] shows the measured angular diameters as a function of observing frequency .
also shown are two curves , one for @xmath53 appropriate for a spectrum with @xmath54 and a diffractive scale @xmath40 much larger than @xmath27 and one for @xmath55 appropriate for the case in which @xmath56 .
as figure [ fig : images ] demonstrates , is not only not a point source , but the two components of it can be resolved into multiple sub - components at high frequencies .
a visual inspection of figure [ fig : diameter ] suggests that at frequencies below 1 ghz scattering dominates the structure of while at frequencies above 1 ghz there are clear deviations from either curve due presumably to intrinsic structure . in order to quantify this result
, we have fit the measured diameters of the eastern component ( table [ tab : models ] ) with the following functional form @xmath57 here @xmath58 is the intrinsic diameter and @xmath59 is the scattering diameter , both at the fiducial frequency of 1 ghz .
unlike in equation ( [ eqn : thd ] ) where we specified the dependence for the specific case of @xmath33 , here we allow the frequency dependence to be part of the fit .
the frequency dependence we use is appropriate for @xmath52 .
several points are in order regarding the choice of data we used . at 0.33 ghz
only one component is seen ( figure [ fig : images]_a _ ) . attempts to fit the visibility data with a two - component structure suggested by the 0.61 ghz image produced a poorer fit to the data .
we have subtracted a one - component model ( table [ tab : models ] ) from the visibility data and imaged the residuals .
there is no indication of a second component at the location expected from the higher frequency images .
we have also produced model visibility data , comprised of two components whose flux densities , diameters , and positions are consistent with those seen at higher frequencies .
fitting these model data with a one - component model results in a fit nearly the same as what we find by fitting the observed visibility data .
the similarity suggests that any systematic error in determining the diameter is probably no more than 1 mas .
because of the self - calibration we performed , we have no absolute position information . given that the eastern component is stronger at 0.61 ghz
, we have therefore assumed that the single component seen at 0.33 ghz is the eastern component . at the higher frequencies
both components can be resolved into multiple sub - components . in figure
[ fig : diameter ] we plot the diameters of the individual sub - components .
the diameters of the individual sub - components are clearly not representative of the diameter of the entire component . therefore , as a measure of the size of the components at these frequencies , we have added the diameters of the sub - components ( table [ tab : models ] ) in quadrature .
if anything , this estimate will _ underestimate _ the size of a component .
however , our primary interest is not in the intrinsic structure , and this procedure suffices to obtain an estimate for the intrinsic size .
is classified as a gigahertz - peaked spectrum source , and both components show a single flux density maximum near 1 ghz @xcite .
accordingly , in equation ( [ eqn : fit ] ) we have assumed that the intrinsic structure has no frequency dependence . while we have not conducted extensive tests , assuming no frequency dependence produces clearly much better fits
than does assuming a frequency dependence of @xmath60 , as might be expected for a homogeneous source with a peak brightness temperature @xmath61 @xcite . we have used a grid - search technique to minimize the difference , in a @xmath62 sense , between the data and the expression of equation ( [ eqn : fit ] ) for the three parameters @xmath58 , @xmath59 , and @xmath32 .
we have also determined the allowed ranges for the parameters in the following manner : holding two parameters fixed at the values that minimize @xmath62 , we change the third parameter until @xmath62 changes by unity .
table [ tab : fit ] summarizes the best fit value for the parameters and their allowed ranges .
the best - fitting curve is also shown in figure [ fig : diameter ] .
a key result of this fit is that a density spectrum spectral index of 4 is highly unlikely .
this result means that the inner scale @xmath27 is substantially smaller than most baselines in the array or that @xmath63 .
the typical array baseline is roughly 2000 km , so we place an upper limit of @xmath64 km .
this constraint is not a serious challenge to the value estimated by @xcite of @xmath65 km . using equation ( [ eqn : thd ] )
we can also derive the scattering measure .
table [ tab : fit ] also lists the values of these derived quantities .
ccc @xmath58 & 3.9 mas & 3.54.3 mas + @xmath59 & 7.0 mas & 6.987.03 mas + @xmath32 & 3.52 & 3.513.53 + sm & @xmath66 + @xmath27 & @xmath67 km + the anisotropy and orientation of a scattered image can , in principle , yield information about the density structures responsible for the angular broadening ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the simultaneous low frequency observations we report here are seemingly ideal for constraining any wavelength - dependent changes in the anisotropy or orientation of , as the @xmath68-@xmath69 plane coverage should be identical ( once scaled by the ratio of the wavelengths ) .
we consider first the anisotropy of the image .
we find axial ratios of @xmath70 ( table [ tab : models ] ) or , in terms of the image anisotropy parameter of @xcite , @xmath71 .
@xcite also provide an estimate of the image anisotropy to be expected from random fluctuations within the medium . for our observing wavelengths and scattering parameters derived above ,
the expected random contribution to @xmath72 is approximately 0.03 .
we conclude that these image anisotropies are significant .
these anisotropies are also comparable to those of other sources seen through the cygnus region @xcite .
@xcite attribute the similarity of the axial ratios , between different sources at different frequencies and different epochs , to be indicative of an anisotropy in the scattering material itself .
we can not conclude that the image anisotropy changes with frequency for .
the formal uncertainties from the fitting of the visibility data are approximately 1% .
even this small uncertainty on the angular diameters , when propagated to the axial ratios , implies an uncertainty of @xmath73 , comparable to the difference between the axial ratios at the two frequencies ( table [ tab : models ] ) .
changes in the image orientation between the two frequencies may reflect changes in the orientation of the density irregularities responsible for the scattering .
unfortunately , we also can not draw any conclusions about the image orientation .
although the @xmath68-@xmath69 plane coverage should be similar , differences in the editing ( due to the significant levels of rfi encountered ) change the relative @xmath68-@xmath69 plane coverage significantly .
an indication of the difference can be seen by inspecting the synthesized beams shown in figures [ fig : images]_a _ and [ fig : images]_b_. in particular , the synthesized beam at 610 mhz has an orientation of 14compared to an orientation for the eastern component of 24 .
thus , the apparent change in orientation between 330 mhz ( @xmath74 ) and 610 mhz ( @xmath75 ) may be due largely to the differences in @xmath68-@xmath69 coverage .
large scale density fluctuations may affect the image position and flux density . in this section
we place constraints on the strength of refractive modulation of the image , focussing primarily on the 0.61 mhz image .
density fluctuations on a scale @xmath76 , where @xmath77 is the characteristic distance to the scattering medium , will cause refractive `` bending '' and `` steering '' of the propagating rays . in turn
, this will affect the angle of arrival of rays and make radio sources wander about a true or nominal position .
the magnitude of this effect depends both upon the strength of scattering and the value of @xmath32 .
@xcite selected because its two component structure allowed for _ relative _ position measurements to be performed , obviating the need for determining absolute positions with milliarcsecond accuracy .
@xcite summarized the predictions for the amount of differential angular wander as a function of @xmath32 as predicted by various groups ( @xcite ; see appendix a of @xcite ) .
they found that the rms differential refractive angular wander is @xmath78 here @xmath79 is the separation between pierce points through the scattering medium . for a double source ( like ) with a component angular separation of @xmath80 , @xmath81 .
the functions @xmath82 and @xmath83 are normalization functions that differ from group to group , but typically by no more than a factor of a few . for reference ,
@xcite predict @xmath84 for a kolmogorov spectrum in a screen located @xmath85 kpc distant , producing an angular broadening of @xmath39 mas , and observed at a wavelength of @xmath86 cm . in general if @xmath87 refractive effects like angular wander dominate over diffractive effects like angular broadening .
thus one would expect to see the apparent position of a source wander by more than the width of its scattering disk . if @xmath52 , the converse is true .
table [ tab : sep ] summarizes the 0.61 mhz measurements of the angular separation between the two components of .
we have estimated @xmath88 using two different methods .
henceforth we shall also distinguish between the ensemble average expected differential angular wander , @xmath88 , and the quantity estimated from the separation measurements , @xmath89 .
first , we calculated @xmath89 using the standard estimator for the variance and using the three values in table [ tab : sep ] .
this estimate is @xmath90 mas .
second , we estimated @xmath89 to be comparable to the maximum difference between component separations at the different epochs , @xmath91 . this estimate is @xmath92 mas .
lcc 1984 august & 60 @xmath93 3 & @xcite + 1988 june & 56 @xmath93 2 & @xcite + 2000 july & 56.1 @xmath93 1.0 & this work + @xmath89 & 2.3 mas + @xmath91 & 4 mas figure [ fig : wander ] shows @xmath88 as a function of the density spectral index @xmath32 , as predicted by @xcite , @xcite , and @xcite .
we also plot our observational limits @xmath89 , @xmath94 , and @xmath91 . in producing figure [ fig :
wander ] we have assumed @xmath95 kpc . we have considered @xmath88 as a function of both @xmath77 and @xmath32 . however , the dependence on @xmath77 is fairly weak , @xmath96 for @xmath54 .
reasonable ranges for @xmath77 produce little difference in @xmath88 .
the value @xmath97 kpc is motivated by the distance to the cygnus superbubble @xcite , through which the line of sight to passes . figure [ fig : wander ] shows clearly that along this line of sight the upper limits on @xmath88 constrain @xmath52 . the largest allowed value , @xmath98 occurs if we adopt an upper limit of @xmath94 and use the expression for @xmath88 from @xcite . using a more stringent limit on the differential angular wander , @xmath91 , the expressions for @xmath88 from @xcite or @xcite we find @xmath99 .
we adopt @xmath33 , as a limit based on @xmath89 we consider to be overly strict .
the standard estimator for the variance , which we have used in determining @xmath89 , assumes that the measurements are independent . in this case
the independence of the various measurements can be assessed by comparing the interval between the different measurement epochs with the refractive time scale , @xmath100 where @xmath69 is the bulk velocity of the material .
@xcite assumed @xmath101100 , for which @xmath1024 yr . in this case , all three measurement epochs ( table [ tab : sep ] ) could be considered independent , though the first two only marginally so .
however , if the characteristic velocity of the scattering medium is smaller , by as little as a factor of 2 , the refractive time scale would be long enough that the first two epochs ( 1984 august and 1988 june ) could not be considered independent .
if the characteristic velocity is as small as 10 , then less than one refractive time has elapsed since the first observation and none of these observations can be considered to be independent .
even if this was the case , we would expect to see some change over the 12 yr time span separating the second and third epochs ( 1988 june and 2000 july ) . that essentially no change in the angular separation is seen
indicates that @xmath91 provides a reasonably stringent upper limit on @xmath88 . in concert with refractive wandering
, the large - scale density fluctuations will also produce intensity fluctuations as more or fewer rays are `` steered '' into the line of sight . in the case of
, this will produce independent variations in the flux density of its two components .
@xcite compared the flux densities of the two components between their observations and those of @xcite .
they found that the flux density of the components was consistent with no change in the 4 yr between 1984 august and 1988 june .
table [ tab : flux ] summarizes these earlier measurements and those reported here .
lcccc 1984 august & 3.8 & 0.6 & 0.16 & @xcite + 1988 june & 3.9 @xmath93 0.1 & 0.6 @xmath93 0.1 & 0.15 @xmath93 0.03 & @xcite + 2000 july & 2.864 @xmath93 0.006 & 0.374 @xmath93 0.014 & 0.13 @xmath93 0.01 & this work + because of changes in telescopes , receivers , correlators , and possible intrinsic changes , we do not compare the absolute flux densities of these earlier epochs to our measurements .
rather we compare the flux density ratio between the two components . as table [
tab : flux ] shows , the western component has remained at a flux density of approximately 15% that of the eastern component for the past 15 yr .
@xcite present expressions for the rms refractive intensity fluctuation @xmath103 as a function of density spectral index @xmath32 ( see also * ? ? ?
* ) . in general ,
@xmath104% for @xmath52 and @xmath105% for @xmath87 .
we determine an observed value @xmath106% from the flux density ratios in table [ tab : flux ] , suggesting @xmath52 . for @xmath52 , the strong scattering ,
i.e. , large sm , determined from the observed scattering diameter ( table [ tab : fit ] ) means that the rms refractive intensity fluctuation will be only a slowly - changing function of @xmath32 . from our observed value of @xmath103 we find ( * ? ? ?
* equation 3.1.5 ) @xmath107 , consistent with the value found by fitting the angular diameters ( table [ tab : fit ] ) .
the pulsar ( 87 from ) has been the subject of many monitoring programs designed at least in part to measure refractive scattering along the line of sight to it . in all cases
, it is found that diffractive scattering dominates refractive scattering , as we find for the line of sight to .
@xcite used dynamic spectra obtained at 0.41 mhz to measure the frequency drift rate and estimated that the diffractive scattering angle was at least 5 times that of the refractive scattering angle . using a structure function analysis of a 0.43 ghz flux density monitoring program , @xcite find diffractive scattering to dominate over refractive scattering .
@xcite monitored the pulsar s flux density at 0.33 mhz and estimated the diffractive and refractive scattering angles toward this pulsar .
they found that the diffractive scattering angle was roughly 10 times larger than the refractive scattering angle .
they also estimated a density spectrum spectral index of @xmath108 .
@xcite were unable to observe any refractive modulations of the pulsar s flux density in their multi - frequency monitoring program , consistent with diffractive scattering dominating on this line of sight .
a key assumption in our analysis of the differential image wander is that there is little intrinsic change in the separation of the components . given that on milliarcsecond scales , many sources show evidence of jets and bulk motion , this assumption clearly must be justified .
we have compared our 1.67 ghz observations with those of @xcite .
this comparison is somewhat difficult because , in general , different antennas at different locations were used in the two vlbi arrays .
thus , this comparison is necessarily crude , but it will be sufficient to show that our assumption is reasonable .
@xcite cite a separation of 60 mas between the two components from observations in 1979 september .
if we fit our 1.67 ghz observations with only two components , we find a separation of 58 mas , _ less _ than that of @xcite . even allowing for differences in arrays and uncertainties in the fitting , we conclude that any intrinsic change in the separation between the two components is probably less than 1 mas . over the 20.8 yr separating these measurements ,
any intrinsic motion is probably less than 0.05 mas yr@xmath109 . at a redshift of 0.354 for @xcite and for an assumed hubble constant of 65 km s@xmath109 mpc@xmath109
, this proper motion corresponds to a projected linear separation velocity of @xmath0 .
our limit is similar to that found by @xcite for , a source with a similar compact double structure .
we have conducted a multi - frequency suite of observations on the source .
we have used these observations to confirm and extend previous conclusions , principally those of @xcite , regarding interstellar scattering along this line of sight .
our 0.61 mhz observations form the third epoch of vlbi observations at this frequency .
our observations are separated by 12 to 16 years from the previous two epochs . as do previous observations
, we find the structure of at this frequency to be dominated by two components approximately 56 mas apart . within the uncertainties of the various measurements , this separation has remained unchanged for the past 16 years .
we can place a reasonably strict limit of 4 mas on any differential image wander caused by refractive interstellar scattering . in turn , this limit implies that the electron density power spectrum along this line of sight has a spectral index near the kolmogorov value , with a value of 4 being highly unlikely . as the diffractive scattering diameter of the source at this frequency is 21 mas , we conclude that diffractive scattering dominates along this line of sight .
the most significant source of uncertainty in this limit is probably the velocity at which the scattering material is drifting past the line of sight .
a `` high '' velocity of 100 would mean that the separations measured at the three epochs are independent .
a `` low '' velocity of 10 would mean that none of the epochs are independent . even if the velocity of the scattering material is low , the separation between our observations and the previous ones is probably long enough that , if refractive scattering were important on this line of sight , some change in the apparent separation of the components would be expected . nonetheless , future observations of at 0.61 ghz or similar observations of other sources on lines of sight for which the refractive time scale could be estimated would be useful .
our observations are in good agreement with a variety of observations on the pulsar ( line of sight 87 from ) .
these observations suggest that diffractive scattering also dominates refractive scattering along the line of sight to the pulsar .
our quasi - simultaneous observations range from 0.33 to 8.4 ghz .
we have used these multi - frequency measurements to constrain the frequency dependence of the angular diameter of one component of . at frequencies below 1 ghz ,
the structure of the eastern component of is dominated by scattering .
fitting for both the density spectrum spectral index @xmath32 and the scattering diameter @xmath39 , we find @xmath110 with the value of 4 again highly unlikely . at a nominal frequency of 1 ghz
, we find the scattering diameter to be @xmath111 mas .
we also place a conservative upper limit on the density spectrum inner scale of @xmath112 km .
we also compared our 1.67 ghz observations to those of @xcite , acquired approximately 21 years previously .
we find no evidence for any change in the separation at this frequency .
this places a limit on the projected linear separation velocity of the two components of @xmath0 .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
basic research in radio astronomy at the nrl is supported by the office of naval research . | we present multi - frequency observations with the very long baseline array of the compact double radio source .
our observations are at 0.33 , 0.61 , 1.67 , 2.3 , and 8.4 ghz , with the 0.61 ghz observations forming the third epoch of observation of this source at that frequency . at 0.61 ghz ,
the structure of is dominated by two components 56 mas apart . within the uncertainties of the various measurements ,
this separation has remained unchanged for the past 16 years . any differential
image wander caused by refractive interstellar scattering is less than 4 mas .
both the lack of differential image wander and the frequency dependence of the angular diameter of below 1 ghz indicate that the electron density power spectrum along this line of sight has a spectral index near the kolmogorov value , with a value of 4 being highly unlikely .
we conclude that diffractive scattering dominates along this line of sight ; results in the literature indicate that this conclusion also holds true for the line of sight to the pulsar ( 87 from ) .
comparison of our 1.67 ghz observations with those obtained 21 years previously place a limit on the projected linear separation velocity of the two components of @xmath0 . |
we construct broad - band seds ( uv to 24@xmath0 m ; fig .
[ fig1 ] ) for 48 radio sources at 1.5ghz with optical spectra in the akari - nep field ( [ 1 ] ) .
following @xcite , we fit an agn and a starburst component additively to each sed . the fractional contribution and luminosities of both components
are derived .
+ + we see a trend for decreasing contribution of active nuclei with increasing radio luminosity ( 3@xmath1 difference between lowest and highest luminosity bins ; fig .
[ fig1 ] ) .
the most radio - loud systems show hints for lower star - formation activity than otherwise expected . | we study the host galaxy properties of radio sources in the akari - north ecliptic pole ( nep ) field , using an ensemble of multi - wavelength datasets .
we identify both radio - loud and radio - quiet agn and study their host galaxy properties by means of sed fitting .
we investigate the relative importance of nuclear and star - formation activity in radio - agn and assess the role of radio - agn as efficient quenchers of star - formation in their host galaxies . |
studies of galaxy formation and evolution at high redshift greatly benefit from multi - object spectroscopy , which is widely accessible from the ground . how far in lookback time one can go is ,
though , severely limited by the emission spectrum of the night sky .
this can be overcome with observations from the space , by performing multi - object spectroscopy which turns out to be rather sensitive to compact sources .
indeed , as shown by the grapes program ( pirzkal et al .
2004 ) , grism spectroscopy with the hst advanced camera for surveys ( acs ) can detect the continuum emission of sources as faint as z@xmath2 27 in 40 orbits .
multi - object spectroscopy with hst comes with low spectral resolution to compensate for the telescope size , and without slits .
as we will discuss in the next sections , the absence of a slit allows spectra of different sources to overlap , the size of the source to decrease the instrumental spectral resolution and the background to increase since each pixel collects the background emission integrated over the whole spectral range of the dispersing element .
slitless spectroscopy has been made available with the hubble space telescope ( hst ) since its launch , in 1990 .
the first instruments to be equipped with grism / prism elements were the wide field planetary camera 1 ( wfpc1 ) and the esa faint object camera ( foc ) .
wfpc1 contained an uv grism ( g200l , horne & mackenty 1991 ) , covering the spectral range between 1300 and 4000 with a resolving power of about 100 ( at @xmath3 = 3000 ) in the first order . the grism was not extensively calibrated in wavelength and flux and was never used for go science observations . during the time of wfpc1
, hst was not yet corrected for spherical aberration .
two prisms were mounted in the foc ; they covered the spectral ranges 1150 @xmath4 6000 and 1600 @xmath4 6000 with a resolving power r @xmath5 100 at @xmath3 = 1500 and @xmath5 250 at @xmath3 = 2500 , respectively .
they provided slitless , fuv and nuv spectroscopy across the f/96 ( 14@xmath6 14@xmath7 ) and f/48 ( 28@xmath6 28@xmath7 ) fields of view .
both prisms were supported modes of foc , i.e. they were calibrated in wavelength and flux before launch and in - orbit ( cf .
paresce & greenfield 1989 , hack 1992 , hack 1996 ) .
nevertheless , the foc prisms were seldomly requested by users during the 7 year life - time of this instrument .
the foc prisms were used to address a number of scientific cases , such as the measurement of the uv flux shortward of @xmath3 = 1800 for a sample of qsos at @xmath8 3 ( jakobsen et al .
1993 ) ; the study of the uv morphology and expansion of the sn 1987a ejecta ( jansen & jakobsen 2001 ) ; and to spatially resolve the spectra of the giant stars in the capella binary system ( young & dupree 2002 ) . since
1997 slitless spectroscopy has become feasible also at infrared wavelengths with nicmos camera 3 , which is equipped with three grisms ( g096 , g141 and g206 ) covering the spectral range 0.8 - 2.5 @xmath1 m at a resolving power of @xmath5 200 ( noll et al .
while the first two grisms exploit the low background of hst , g206 is subject to the thermal background emission from hst .
the three grisms provide multi - object spectroscopy over a field of view of 51@xmath6 51@xmath7 at a spatial resolution of 0.2@xmath7/pix , and were mainly used to study the evolution of field galaxies at 0.7 @xmath9 1.9 ( mccarthy et al .
1999 ) and to investigate the nature of planets and kuiper belt objects ( mccarthy et al .
1998 ) . during the first two years of nicmos ,
the grisms were calibrated in wavelength and flux in orbit , by observing a galactic planetary nebula and a white dwarf .
software for extracting spectra ( nicmoslook ) was released by freudling & pirzkal ( 1997 ) and freudling ( 1999 ) .
the nicmos grisms have been quite extensively used , by both go proposals and parallel programs . prior to the servicing mission 3b ,
the accepted go proposals requiring spectroscopy represented about 14@xmath10 of all the successful nicmos proposals , while in the case of foc the approved proposals making use of the prisms amount to only @xmath5 4@xmath10 of the total . although designed as a slit , high - to - medium resolution spectrograph ,
stis has also provided slitless spectroscopy mostly with the g750l grating ( r @xmath5 750 at 7500 ) when employed in parallel mode .
an iraf routine ( slitless in stsdas / stecf ) was developed by n. pirzkal for the extraction of slitless spectra . the slitless parallel data of stis have been analyzed to search for new low - mass stars and brown dwarfs ( plait 1998a and 1998b ) , to study the intermediate - redshift universe ( de mello & pasquali 2001 , teplitz et al . 2001 , teplitz et al .
2003 ) and to derive the hydrogen reionization edge ( windhorst et al .
2000 ) . in march 2002 ,
hst was refurnished with a new instrument , the advanced camera for surveys ( acs ) which offers the unique combination of a large field of view ( [email protected] @xmath12 [email protected] ) with a high angular resolution ( 0.05@xmath7/pix , cf .
the main task of acs is to perform deep imaging and spectroscopic surveys .
in contrast to older generation instruments , the acs grism and prisms have been extensively calibrated in laboratory and specific go calibration proposals have been set up to assess their in - orbit performance .
it is believed that such an effort will involve a larger number of hst users in using the acs spectroscopic capabilities and will return relevant scientific results . in the first two years of acs
, the grism has been extensively used for the search for supernovae type ia at high redshift ( magee et al . 2002 , tsvetanov et al .
2002 , blakeslee et al .
2003 , riess et al . 2004 ) and for high - redshift ly@xmath13 emitters , and to study galaxy formation and evolution at intermediate - to - high redshifts , especially in the udf field ( pirzkal et al .
2004,2005a , b ; xu et al . 2005 ; daddi et al .
2005 ; malhotra et al .
2005 ; pasquali et al . 2005a , b ; rhoads et al .
2005 ) . in this paper
, we review the concepts and technique of the grism calibrations , which have been developed for non - drizzled , geometrically distorted data . drizzling the grism data ,
_ before establishing their dispersion solution
_ , involve a change of pixel size and geometry which in turn affect the dispersion and zero point of the wavelength calibration and also the amplitude and wavelength dependence of the flat - field correction for grism spectra .
these quantities depend on the drizzle parameters which can be specified differently each time by different users .
for these reasons , the grism calibration has been performed only for raw grism data .
the acs is a three - channel instrument which performs broad and narrow band filter imaging from uv to nir , polarimetric imaging at ultraviolet and optical wavelengths , and low - resolution slitless spectroscopy in the range from @xmath5 1150 to @xmath3 @xmath5 1 @xmath1 m with a set of grism and prisms ( cf .
http://www.stsci.edu/hst/acs/documents ) .
the acs spectral elements can be summarized as follow : the wide field channel ( wfc ) has a [email protected] @xmath12 [email protected] field of view at a spatial resolution of 0.05@xmath7/pix and is equipped with a grism ( g800l ) covering the spectral range from 5500 to 1 @xmath1 m with a response peak at @xmath5 7200 and @xmath5 6000 in the 1@xmath14 and 2@xmath15 orders , respectively .
the grism resolving power is @xmath5 100 and the dispersion is nearly linear , @xmath5 39 / pix in the 1@xmath14 order and @xmath5 20 / pix in the 2@xmath15 order .
the tilt of the spectra is nearly -2@xmath16 from the image x axis , varying slightly across the field .
the high resolution channel ( hrc ) is characterized by a 26@xmath6 29@xmath7 fov with a spatial resolution of 0.027@xmath7/pix and uses the same grism of the wfc , which introduces a tilt in the spectra of -38@xmath16 ( from the image x axis ) and offers a dispersion of @xmath5 24 / pix in the 1@xmath14 order and of @xmath5 13 / pix in the 2@xmath15 order .
the hrc also features a prism ( pr200l ) covering the 1600 - 5000 range , with a non linear dispersion varying from 2.6 / pix at 1600 to 105 / pix at 3500 and 560 / pix at 5000 . the solar blind channel ( sbc ) has a field of view of 31@xmath6 35@xmath7 where each pixel is 0.032@xmath7 in size .
two prisms are available with the sbc : pr110l covering from 1150 to 2000 , and pr130l from 1220 to 2000 (i.e .
geocoronal ly@xmath13 is blocked ) .
the dispersion of pr110l is 2.6 / pix at 1216 , while the dispersion of pr130l varies from 1.65 / pix at 1250 to 20.2 / pix at 1800 .
the acs grism and prisms are used in slitless mode .
this raises a number of important issues : _ i ) _ the extension of a source along the dispersion axis is given by the spatial profile of the object , and affects the achievable spectral resolution . _
ii ) _ the zero point of the grism wavelength calibration is defined by the position of the object in the direct image . when no direct image of the source is available , the position of the 0@xmath17 order can also be used ,
though in a less accurate way since the 0@xmath17 order is itself dispersed .
_ iii ) _ the spectroscopic flat - field is field and wavelength dependent . depending on the position of a target in the direct image , the same pixel in the grism / prism image receives different wavelengths . _
iv ) _ each pixel in a grism / prism spectrum has a contribution from the background emission which is integrated over the entire spectral range of the grism / prism .
_ v ) _ depending on the luminosity of a source , the grism orders higher than 1 ( at higher x coordinates in the grism image ) and the grism negative orders ( at lower x coordinates ) can be identified .
no filter can be used to suppress higher and negative orders . _
vi ) _ depending on the position of the object in the direct image , the source spectrum can be truncated in the grism / prism image .
objects outside the field of view can still produce spectra in the grism / prism image . _ vii ) _ the extension of grism spectra and the number of their orders can give overlap and contamination among spectra of adjacent objects .
therefore , the wavelength calibration of the acs grism and prisms requires a full description of the grism / prisms physical properties , such as the tilt of the spectra , the length of the prism spectra and of the grism orders together with their separation from the grism 0@xmath17 order , the ( x , y ) offsets between the object position in the direct image and the corresponding 0@xmath17 order in the grism image . indeed
, these quantities allow the identification of the spectrum in the grism / prism image of any object in the direct image , the tracing of spectra and set the size of the aperture extraction for each prism spectrum and grism order ( pasquali et al .
2003a , b ) . as described in the next sections
, these parameters are field - dependent because of the geometric distortions of acs .
this is also true for the flat - field , which an accurate flux calibration of the spectra depends on .
since no slit is available , the nominal dispersion of the acs grism gets convolved with the object size and , in the case of non - round sources , with the object size along the dispersion axis as determined by the object orientation .
this convolution obviously lowers the grism resolution and plays a key role in planning the spectroscopic calibrations and the go observations with the acs ( pasquali et al .
2001a ) .
to quantify the effect of object size and orientation on the effective dispersion of the grism , we have simulated the optical spectrum of the galactic planetary nebula ngc 7009 using slim 1.0 ( pirzkal & pasquali 2001 , pirzkal et al .
slim 1.0 is a slitless spectroscopy simulator developed by n. pirzkal and is designed to produce realistic and photometrically correct spectra acquired with the acs grism and prisms .
briefly , any input spectrum is redshifted as required and scaled to a desired magnitude in a reference filter , then dispersed according to the grism / prisms dispersion solution . at each wavelength , the dispersed spectrum is convolved with the grism / prisms throughput curve and with a gaussian psf whose fwhm value is set by the hst diffraction limit at any specific wavelength .
two sets of simulations have been produced for both the wfc and the hrc which do not include background or read - out noise .
the first set is based on the assumption that ngc 7009 is disc - shaped , with uniform surface brightness and diameter varying from [email protected] to [email protected] , [email protected] and 1@xmath7 for the wfc , and from [email protected] to [email protected] and [email protected] for the hrc .
the second set assumes that ngc 7009 is elliptical with a varying position angle ( pa ) with respect to the dispersion axis .
the nebula size is [email protected] @xmath12 [email protected] and [email protected] @xmath12 [email protected] for the wfc and the hrc respectively .
the pa of the nebula major axis is assumed to vary from 0@xmath16 to 20@xmath16 , 45@xmath16 and 90@xmath16 so that the object major axis progressively goes from being aligned along to being perpendicular to the dispersion axis .
we have measured the fwhm of a few lines in the 1@xmath14 and 2@xmath15 order spectra obtained for the wfc and the hrc in both sets of simulations , in order to quantify the _ degradation _ of the grism nominal dispersion as a function of object size .
these fwhm values have been normalized to the fwhm of a source with a [email protected] diameter and are plotted in figure 1 as a function of the object size along the dispersion axis in arcsecs , for both the wfc and the hrc .
as expected , the line fwhm nearly doubles by doubling the object linear size , implying that the grism dispersion degrades by almost the same factor as the object size increases .
the flattening in the curves for linear sizes larger than 3 @xmath12 [email protected] for the wfc ( 3 @xmath12 [email protected] for the hrc ) is due to severe line blending which makes the line width measurements unreliable .
during the ground testing of acs at the goddard space flight center , the refractive aberration simulator ( ras / ramp ) was used to acquire a series of grism spectra for both the wfc and the hrc .
specifically , spectra were obtained through a 4 @xmath1 m pinhole aperture fed by an argon lamp and qth fiber lamp sources .
spectra were acquired at different positions across the field of view , located close to the center and to the ccd amplifiers in the corners .
these data provided a first wavelength calibration , and also allowed to check the amplitude of the field dependence of the grism resolution .
the spectra of the continuum lamp were used to derive the relative , integrated throughput of the different grism orders across the field of view , but were not suited to determine the throughput as a function of wavelength for the various grism orders .
we could then establish that the total power in the 0@xmath17 order is 2.5@xmath10 of that in the 1@xmath14 order , and 1@xmath10 and 0.5@xmath10 in the positive and negative higher orders , respectively .
since the grism is sensitive to wavelengths redder than 5500 , we could identify and use only the argon lines redward of 6700 .
this made the wavelength solutions of the grism orders less accurate than those derived in orbit ( cf . sects .
9 and 10 ) .
for example , the wavelength solution of the positive 1@xmath14 order could only be fitted with a first - order polynomial , which , later on , gave raise to significant wavelength offsets in the spectra of the in - orbit calibrators .
the results obtained in sect . 3 pose important constraints on the planning of wavelength calibration observations , such as those foreseen for the servicing mission orbital verification ( smov , following the installation of acs onboard hst ) tests and subsequent go calibration programs .
acs is not equipped with arc lamps , therefore the in - orbit dispersion correction of the grism can only be determined from observations of specific astrophysical sources . for this purpose , calibration targets should satisfy the following requirements : _ i ) _ the spatial extension of the target should be minimal , to allow accurate determination of the dispersion and wavelength zero point ; _ ii ) _ the target brightness allows for reasonably short exposure times ; _ iii ) _ the target spectra should contain a significant number of unblended emission lines ; _ iv ) _ no extended nebulosity is associated with the targets which would degrade the dispersion of the instrument and increase the local background ; _ v ) _ the spectrophotometric variability ( either intrinsic or induced by an eclipsing companion or by orbital motion ) is negligible so that emission features can be identified at the same wavelength at any observational date ; _ vi ) _ the targets should not lie in crowded fields to avoid contamination by nearby spectra .
spectroscopy in slitless mode relies on a pair of direct and grism images , since the zero point of the grism wavelength solution is set by the position of the object in the direct image . because the acs is tilted with respect to the optical axis of hst , its field of view suffers severe geometric distortions which introduce a field dependence in the grism dispersion and wavelength zero point .
hence , calibration targets have to be observed at a number of positions across the wfc and hrc chips in order to be able to parameterize the field dependence of the grism wavelength solution .
this need strengthens requirement _ i ) _ : the exposure times for the calibration targets should be reasonably short to allow multiple acquisitions across the field of view .
in what follows , we discuss the observational strategy adopted for the smov tests and following go calibrations in order to achieve the wavelength calibration of the acs grism . at optical wavelengths ,
the spectra of planetary nebulae are dominated by strong , narrow emission lines mainly due to h , hei , [ o ii ] , [ n ii ] , [ o iii ] and [ s iii ] .
the [ o iii ] and [ nii ] lines are often used to measure the expansion velocity of planetary nebulae ( pne ) ; galactic pne span a wide range in expansion velocities , from @xmath18 5 km s@xmath19 to 50 km s@xmath19 ( acker et al .
1992 ) . in the worst case
, an expansion velocity of 50 km s@xmath19 corresponds to a fwhm value of @xmath18 1 at the restframe h@xmath13 line . with a resolving power of r
@xmath5 100 , the acs grism does not resolve any planetary nebula in velocity and any planetary nebula would be equivalent to a ground arc lamp . among all the known pne ,
only those which are compact ( with a full size smaller than 0.1@xmath7 ) can be observed with the acs grism according to the requirements listed in sect .
4 . pne in the galaxy and in the magellanic clouds are rarely compact when imaged with hst and even when compact often have a low intensity halo .
therefore , in order to preserve the nominal dispersion of the acs grism , we have to resort to more distant pne , for example those in m31 .
to assess the feasibility of grism observations of extragalactic pne in terms of exposure time and s / n ratio in the lines , we simulated the acs spectra of the brightest and most compact pne in m31 using slim 1.0 ( cf .
pasquali et al .
the selected pne were originally studied by ciardullo et al .
( 1989 ) and jacoby & ciardullo ( 1999 ) .
our simulations show that pne can be used to determine the dispersion correction of the grism 1@xmath14 order , but fail the wavelength calibration of the 2@xmath15 since only two lines can be detected with a reasonable s / n ratio . to improve this , the integration time should be longer by a factor of @xmath183 ( equivalent to @xmath18 1 h per pointing ) , thus making pne too time - consuming for routine go calibration programs . an additional disadvantage of observing pne in m31
is the field crowding , which makes spectra of adjacent sources overlap .
nevertheless , pne can be employed as secondary targets , to verify the accurateness of the wavelength calibration .
having discharged pne , our simulations show that galactic wolf - rayet ( wr ) stars satisfy the calibration requirements . as figure 2
points out , wr stars of spectral type wc6 - wc9 have a large number of bright emission lines ( he and c ) in the spectral range covered by the acs grism . the only drawback here is that the velocity of their stellar winds considerably broadens their emission features .
the known galactic wr stars ( van der hucht 2001 ) have wind velocities between @xmath18 700 km s@xmath19 and 3300 km s@xmath19 with a mean wind speed of about 1700 km s@xmath19 and with 19@xmath10 of the sample having a wind velocity larger than 2100 km s@xmath19 .
we thus assumed a wind speed of 2000 km s@xmath19 and computed the line broadening induced in wfc and hrc grism spectra of wr stars .
the results indicate that such a stellar wind is hardly resolved in the 1@xmath14 order spectra and partially affects the 2@xmath15 order spectra .
higher wind velocities rapidly decrease the grism resolution at any order . clearly , the wind velocity constraint ( @xmath20 2000 km s@xmath19 ) has to be added to the selection criteria discussed earlier in sect .
4 , together with the requirement that the variability of the selected targets is negligible .
spectro - photometric variability could change the number of detectable emission lines , and variability in the stellar wind could change the line fwhm ; both cases would make the grism wavelength calibration less accurate . from the catalogue of van der hucht ( 2001 ) ,
we have selected two stars , wr45 and wr96 , whose spectral type , coordinates , v magnitude and wind velocity are listed in table 1 .
we produced slim 1.0 spectra based on two wr templates kindly provided by p. crowther and closely matching the spectral types of wr45 and wr96 .
the simulations are plotted in figures 2 and 3 for the wfc and the hrc 1@xmath14 and 2@xmath15 orders , respectively . in the case of the wfc simulations , an exposure time of 10
s and 60 s has been assumed for the 1@xmath14 and 2@xmath15 order , respectively , while for the hrc 20 s and 60 s. contrary to the planetary nebula case , the simulated spectra of wr45 and wr96 indicate that a relatively large sample of emission lines can be used for the wavelength calibration of both the 1@xmath14 and 2@xmath15 orders of the acs grism . as a counter - check , the simulated spectra of wr45 and
wr96 have been used to derive the grism wavelength solution for both the wfc and the hrc .
we have measured the position of each identified line in the slim output in pixels scaled to the position of the object in the direct image , and fitted pixels against wavelength assuming a first order polynomial .
the accuracy of the dispersion correction thus obtained is better than 0.3 pixels for both the 1@xmath14 and 2@xmath15 order ( i.e. 12 and 7 , respectively ) . since no high resolution spectra were available in literature at the time of the acs ground calibrations , we observed wr45 and wr96 with the emmi spectrograph , mounted on the eso ntt telescope ( 266.d-5653 , pi pasquali ) .
the observations were performed in service mode in june and august 2001 as part of the eso director general discretionary time .
spectra were acquired through the remd grating @xmath21 8 ( between 5000 and 7500 at 1.26 / pix , and in the range 7300 - 9750 at 1.26 / pix ) and the rild grism @xmath21 2 ( from 4000 to 9000 at 2.7 / pix ) .
the high resolution spectra were used for an accurate identification of the he and c lines typical of wr stars , while the low resolution spectra provided the flux calibration of the stellar continuum and lines .
the ntt , high resolution spectra of wr45 and wr96 are plotted in figure 4 .
they were used as inputs to slim 1.0 in order to simulate the wfc and hrc grism observations of wr45 and wr96 and to optimize the exposure time and s / n value of the in - orbit calibrations .
as part of the smov and interim calibration programs ( 9029 and 9568 , pi pasquali ) , we observed wr45 at a number of positions across the field of view of the wfc and the hrc .
wr96 was also observed , but only with the wfc , at the same positions used for wr45 .
we selected 10 pointings for the wfc ( among which the five from the ground tests ) and 5 ( only those observed during the ground calibrations ) for the hrc , in order to sample the field dependence of the grism properties , which ground - based calibrations indicated to be smaller for the hrc .
wr96 was re - observed in cycle 12 together with the planetary nebula smp-81 in the lmc ( program 10058 , pi walsh ) .
these observations were designed to provide additional sampling points across the field of view to better map the spatial variation of the wavelength solution .
the interim , and cycle 12 data ( i.e. the object positions in the direct image ) are illustrated in figures 5 and 6 . at each pointing a pair of direct and grism images
was acquired and repeated at least twice , either during the same visit or in different visits , to test the repeatability of the filter wheel positioning .
direct images were taken with the f625w and f775w filters to test whether the target position in the direct image is filter - dependent .
exposure times for imaging and spectroscopy are listed in table 2 for both channels and targets . in particular , the spectroscopic exposure times were tuned to get a s / n ratio @xmath22 30 in both the positive and negative 1@xmath14 and 2@xmath15 orders , and turned out to be long enough to detect the 3@xmath23 and -3@xmath23 orders .
the observations included also two white dwarfs , gd 153 and g191b2b , which were acquired at the same positions across the chips as the wr stars .
their spectra were used to derive the grism throughput and flux calibration of the wfc and the hrc grism configurations .
the spectral tilt can be derived from the ( x , y ) coordinates of the line - emission peaks , as measured along the spectrum from the -3@xmath23 to the + 3@xmath23 order in the grism spectra .
we have fitted up to 20 ( x , y ) pairs with a first order polynomial , whose slope is the tilt of the grism spectra with respect the x axis of the grism image . on average , the spectral tilt in the wfc / grism configuration is [email protected] @xmath24 [email protected] from the image x axis , but it varies by 1@xmath16 across the field of view of the wfc ( cf .
figure 8) .
the maximum variation ( @xmath18 [email protected] ) occurs along the diagonal from the top left corner to the bottom right ; a variation of [email protected] is seen along the diagonal from the bottom left to the top right corner of the field of view ( cf .
pasquali et al .
2003a ) .
the spectra appear off - set from the object in the direct image .
we measured the positions of the 0@xmath17 order and the object in the direct image using their light centroid .
we used the fits of the spectral tilt to calculate the y@xmath11 coordinate on the spectrum corresponding to the x coordinate of the object in the direct image . the difference ( y@xmath25 - y@xmath11 ) represents an offset between the object in the direct image and the grism 0@xmath17 order , which allows to locate the spectrum of a source in the grism image once the source coordinates have been measured in the direct image .
the mean value of this y - offset across the field of view is -3 pixels , although the y - offset is itself field dependent .
the offset of the object in the direct image from the grism 0@xmath17 order along the x axis is almost constant across the wfc field of view and is of about 112 pixels ( cf .
pasquali et al .
2003a and figure 8) . to estimate the length of the grism orders along the image x axis
, a threshold has to be set in the counts which distinguishes between pixels of background and pixels of source spectrum .
we set this threshold to 3@xmath26 level above the background , and measured the x coordinates of the blue and red edges of each order spectrum .
these points give an estimate of the length and separation of the grism orders , which show negligible field dependence .
the 0@xmath17 order turns out to be dispersed over 23 pixels and its fwhm is about 4 pixels .
the 1@xmath14 order is about 156 pixels long , while the 2@xmath15 is @xmath18 125 pixels in length .
it should be about two times more extended than the 1@xmath14 order , but the very low throughput of the grism in the 2@xmath15 order prevents its detection at @xmath27 8000 ( cf .
figure 8) .
the -1@xmath14 and -2@xmath15 orders are 102 and 111 pixels , respectively .
note that there is contamination in the first order spectrum at @xmath27 10000 by the grism second power . during the analysis of the combined interim and cycle 12 data , we used a different method to trace the spectral orders .
the centroid was measured directly on the images in the cross - dispersion direction along the spectral traces , and the run of y - offset versus x position with respect to the target position in the direct image was fitted with a straight line .
we found this method to provide nearly identical results to the approach described above .
the mean spectral tilt across the hrc field of view is [email protected] @xmath24 [email protected] from the image x axis ( cf .
figure 9 ) . the spectral tilt decreases from the top left to the bottom right corner by @xmath18 [email protected] , and by @xmath18 [email protected] along the diagonal from the bottom left to the top right corner of the field of view ( pasquali et al .
2003b ) .
the y - offset is on average -1.5 pixels with little field dependence , while the x - offset between the grism 0@xmath17 order and the object in the direct image is 145 pixels ( as shown in figure 9 ) .
both the grism 1@xmath14 and 2@xmath15 orders are about 195 pixels along the spectral trace , while the -1@xmath14 and -2@xmath15 orders are @xmath18 183 and 174 pixels , respectively . as for the wfc / grism configuration ,
the sensitivity of the 2@xmath15 order redward of 8000 is so low that the 2@xmath15 order spectrum is not detected any longer and turns out to be as extended as the 1@xmath14 . the 0@xmath17 order
is dispersed over 9 pixels along the spectral trace and its fwhm is 6 pixels .
the wavelength calibration of the grism spectra of wr45 and wr96 relies on the set of template spectra taken for these two stars at medium resolution with the eso ntt / emmi spectrograph .
two different approaches were adopted for determining the wavelength solutions .
we first summarize the analysis of the smov and interim data , which is described in more detail in pasquali et al .
( 2003a ) . in this case
, the wfc / hrc grism spectra were extracted with the st - ecf package axe ( pirzkal et al .
2001a , pirzkal et al .
2003a , b ) , adopting an extraction aperture 2 pixels wide ( to sample the peak of the instrument psf ) and the background was estimated 100 pixels away from the spectral trace .
since no wavelength solution had been defined at this stage , the wavelength scale of the extracted spectra was in units of pixels along the spectral trace , relative to the position of the object in the direct image .
seven emission lines can be identified in the extracted @xmath241@xmath14 and 3@xmath23 orders , while three to five lines are identified in the @xmath242@xmath15 and -3@xmath23 orders .
we measured for the wfc a mean line fwhm@xmath28 of 3 , 4 and 6 pixels in the @xmath241@xmath14 , @xmath242@xmath15 and @xmath243@xmath23 orders , and for the hrc 4 and 6 pixels for the @xmath241@xmath14 and 2@xmath15 orders respectively .
the ntt template spectra were convolved with a gaussian function whose @xmath26 was set to fwhm@xmath28/2.36 in order to mimic the acs grism spectra .
we reidentified the emission line in the new `` degraded '' templates , and measure their centroid wavelength via multi - gaussian fitting and deblending . at the same time and with the same procedure
, the line centroids were measured in the wfc / hrc grism spectra in units of pixels .
this method produced a table of line peak wavelengths and pixels for each grism order , which were then fitted to derive the wavelength solution of each spectral order .
clearly , the order of the fit depends on the number of identified lines available and on the s / n ratio in the extracted wfc / hrc grism spectra . for the analysis of the interim and cycle 12 wfc observations of wr96 ( larsen & walsh 2005 ) the grism spectra were again extracted with axe as before , except that an extraction box width of 10 pixels was used .
the wavelength solutions were then determined by fitting the ntt spectra directly to the acs grism spectra . the amoeba routine ( press et al .
1992 ) was used to minimize the r.m.s .
difference between the smoothed ntt spectra and the grism spectra as a function of the smoothing length and the wavelength solution coefficients .
since the acs spectra were not flux calibrated at this stage , we also fitted for a wavelength - dependent sensitivity curve , approximated for this purpose as a 4th order polynomial .
figure 7 shows the smoothed ntt spectrum of wr96 and the acs grism spectrum for the best - fitting wavelength solution . clearly , the fit is not perfect , perhaps partly due to real temporal variations in the spectra of the wr star .
repeated iterations of the amoeba fits were made , each with slightly different initial guesses for the fit parameters to test the repeatability of the fits .
generally , the wavelength zero - points reproduced within a few and the dispersions to within 0.1 pixel@xmath19 .
all these procedures were applied to each grism order for each position across the field of view .
an example of 2d grism images taken with the wfc and the hrc is shown for wr96 in figures 8 and 9 respectively , where the different grism orders have been also identified . in these figures
we compare the location of the grism orders with the object position in the direct image .
the wr45 spectra for the positive and negative orders taken at the center of chip 1 are plotted in figure 10 , in units of counts vs wavelength in . in the case of the wfc / grism observations , six orders ( from the -3@xmath23 to the + 3@xmath23 orders ) were detected together with the 0@xmath17 order .
their wavelength solutions computed for the pointings across the wfc field of view are summarized in tables 3 and 4 .
the blank entries indicate that the order was not detected because it fell outside the physical area of the chips .
these wavelength solutions have been computed from the combined interim and cycle 12 calibration data ( larsen & walsh 2005 ) .
they are reassuringly similar to those obtained from the smov / interim data ( that are omitted , for this reason , from this paper . for more details cf .
pasquali et al .
indeed , the difference between the two calibrations is generally less than one pixel .
thus , users who have been using the older calibrations can consider their wavelength scales accurate to better than about 1 wfc pixel , but for future applications we recommend that the updated calibrations be used .
these will be available via the st - ecf web pages .
the wavelength solution obtained for the 1@xmath14 and 2@xmath15 orders is a second order polynomial in the form of : @xmath3 = @xmath29 + @xmath30x@xmath11 + @xmath31x@xmath32 , where @xmath29 is the wavelength zero point in , @xmath30 the first term of the dispersion in / pix and @xmath31 the second term of the dispersion in / pix@xmath33 .
x@xmath11 is the distance from the x coordinate of the object in the direct image along the spectral trace . for the 1@xmath14 order ,
the first term of the dispersion @xmath30 significantly varies across the field of view , with an amplitude of about 20@xmath10 ( of the value measured at the center of the field ) from the the top , left corner to the bottom , right corner . along this same direction ,
the wavelength zero point and the second term of the dispersion vary by @xmath18 20 and a factor of @xmath18 2 respectively . in the case of the 2@xmath15 order ,
@xmath30 varies by about 10@xmath10 of the dispersion measured at the center of the field , along the diagonal from the top , left corner to the bottom , right corner of the field of view .
along this same direction , the wavelength zero point increases by @xmath18 220 .
the wavelength solutions of the positive 3@xmath23 and all the negative orders are a first order polynomial in the form of : @xmath3 = @xmath29 + @xmath30x@xmath11 .
they exhibit the same field dependence as for the positive 1@xmath14 and 2@xmath15 orders , with @xmath30 varying by about 15@xmath10 from the center to the bottom , right corner of the field of view .
it is worth noticing here that the 3@xmath23 orders are out of focus .
since the focal plane is tilted with respect to the optical axis the higher orders are more displaced from the optical axis .
consequently , a number of lines appear split when compared with the template spectra of wr45 and wr96 ( cf .
figure 11 , where the split lines are indicated with a vertical line ) .
in the hrc grism data we detected six orders together with the 0@xmath17 order , from the -3@xmath23 to the + 3@xmath23 order . because of the position of wr45 across the field of view and the severe tilt of the spectra , the @xmath243@xmath23 orders are truncated and only a pair of emission lines can be identified .
we thus do not provide any wavelength solution for these orders .
the @xmath241@xmath14 and @xmath242@xmath15 order spectra acquired at the center of the field of view are plotted in figure 12 in units of counts vs wavelength in .
the wavelength solutions obtained across the hrc field of view are listed in tables 5 and 6 ( cf .
pasquali et al .
2003b ) .
the wavelength solution of the 1@xmath14 order is fitted with a second order polynomial as in the case of the wfc 1@xmath14 order .
the dispersion is seen varying by 8@xmath10 between the top , left corner to the bottom , right corner of the field of view . in this case
, the field dependence follows the same diagonal as for the wfc , only in the opposite direction .
the wavelength zero point also varies by 0.2@xmath10 between these two edges of the field . a first order polynomial traces the wavelength solution of the grism 2@xmath15 order .
the dispersion varies again by @xmath18 8@xmath10 from the top , left corner to the bottom , right corner of the field , and it is accompanied by a variation of 0.3@xmath10 in the wavelength zero point .
the -1@xmath14 and -2@xmath15 orders are detected only in two positions , at the center and the bottom , right corner of the field .
the flux calibration of the acs grism requires the correction for flat - field and for the total grism throughput , which takes into account the intrinsic grism response , the quantum efficiency of the ccd detectors and the throughput of the telescope and instrument optics . in the case of slitless spectroscopy ,
the flat - field is field and wavelength dependent at the same time . indeed , depending on the coordinates of a source in the direct image , the same pixel in the grism image is exposed to different wavelengths during different telescope pointings and grism acquisitions .
a grism flat - field can be then constructed via interpolation of the flat - fields available for different imaging filters at any pixel position and wavelength , i.e. what is refereed to as a flat - field cube . for this purpose
, we used a 3@xmath23 order polynomial as a function of pixel coordinates and wavelength , which is directly used by the extraction package axe ( pirzkal et al .
2003a , b ) .
technically , the correction for flat - field is applied following the wavelength calibration of a spectrum : only at this point , it is possible to extract a monodimensional flat - field from the flat - field cube in correspondence with the source position in the direct image and the wavelength range of the source spectrum .
ideally , the grism flat - field cube should be constructed from the flat - fields of narrow - band filters , in order to better sample the wavelength space of the grism spectra and therefore to perform an accurate flux calibration .
practically , the available grism flat - field cubes ( for the wfc and the hrc at http://www.stecf.org/instrument/acs ) are based on in - orbit flat - fields for broad - band filters which describe the large scale variations in the ccd response up to @xmath34 9000 . in the case of the wfc , taking the broad - band filter flat - fields at face value generates a grism flat - field cube which produces flux discrepancies among the grism spectra of the standard star gd153 ( the same occurs for g191b2b ) observed at different positions across the field of view .
these differences in flux possibly arise from a large - scale flat properties of the grism images that are different from those seen in direct images .
in addition , since the imaging flat - fields are designed to give a flat image of the sky , their application to a grism image introduces a correction for the geometric distortions which is already been taking into account in the wavelength calibration of the extracted spectra .
consequently , the differences in flux among the standard star spectra taken in different positions were fitted with a 2d surface which was used to generate a _ corrected _ grism flat - field cube , able to produce spectra of the same star which agree within 1@xmath10 between 6000 and 9000 and 2 - 3@xmath10 at redder wavelengths , all across the field of view of the wfc ( pirzkal et al .
2003a , b , walsh & pirzkal 2005 ) . in the case of the hrc ,
the standard star gd153 was observed only in three positions out of the five which were adopted for the wavelength calibration .
a simple grism flat - field cube , without any correction as applied to the wfc cube , generates spectra whose fluxes already agree within 2 @xmath10 , most likely because of the much lower degree of geometric distortions in the hrc with respect to the wfc ( pirzkal et al .
2003b ) .
the spectra extracted for gd153 and g191b2b , calibrated in wavelength and properly corrected for flat - field and ccd gain , were normalized by the exposure time , and divided by the stsdas template spectrum calibrated in flux .
the resulting sensitivity functions ( in e@xmath35 pix@xmath19 per erg @xmath36 @xmath19 ) were averaged into a single grism response curve independent of the position across the chip and the standard deviation of the mean was assumed to be the error on the grism throughput .
the errors in the absolute flux calibration are then 3@xmath10 between 5000 and 9000 and 20@xmath10 past 10000 .
the mean wfc / hrc grism responses obtained for the 1@xmath14 order are plotted in figure 13 .
they are delivered with the axe package and can be found at http://www.stecf.org/software/axe together with the grism higher order responses .
a secondary effect in the wfc / hrc grism spectra is fringing , due to the interference between the incident light and the light reflected at the interfaces between the thin layers of the ccds .
the modeling of fringes is possible once the layer composition of the ccd detector , the thickness and material of these layers are known . in the case of acs , these parameters are not available ( the wfc ccds have proprietary construction ) , and are estimated by comparison of a fringe model with narrow - band filter flat - fields .
walsh et al .
( 2003 ) have found that the peak - to - peak amplitude of the fringes in the hrc ( @xmath5 27@xmath10 ) is reproduced by assuming a ccd model with four layers : the si top layer 12.5 - 16.0 @xmath1 m thick , the second sio@xmath37 layer 0.26 @xmath1 m in thickness , the third si@xmath38n@xmath39 and the fourth si layers with a 0.23 @xmath1 m and 0.83 @xmath1 m thickness respectively .
similarly , the wfc fringes are characterized by a peak - to - peak amplitude of @xmath5 24@xmath10 which is reproduced by the same four layers ccd model adopted for the hrc , only with different thicknesses : 12.6 - 17.0 @xmath1 m for the top layer , 1.14 @xmath1 m , 0.90 @xmath1 m and 3.08 @xmath1 m for the second , third and fourth layer respectively .
the predictions of the above ccd models can be use to correct the narrow - band filter flat - fields for fringing : the net result is a decrease of the fringes peak - to - peak amplitude of about four .
does fringing affect grism spectra ?
the incident light on the grism is convolved with the instrument psf along both the dispersion and cross - dispersed axes . in the case of the wfc ,
the fwhm of the psf is @xmath5 2.5 pixels , or 100 assuming @xmath30 = 40 .
the fringes period is @xmath1870 , therefore the psf has the net effect of smoothing out the fringes envelope , by a factor of six as slim simulations show .
therefore , fringing in the wfc grism spectra is practically negligible . since the grism dispersion is higher when coupled with the hrc ( @xmath30 @xmath5 24 ) and the fringes period is 65 , some residual fringing
may still be observed in hrc grism spectra ( walsh et al .
2003 ) . at the present time
, a routine for the fringing correction is being planned for axe .
the in - orbit calibrations of the acs grism served the purpose to characterize its physical properties ( i.e. spectral tilt ) and the wavelength solution for each of its detected orders .
equally important , the calibrations were performed in selected positions across the wfc and hrc fields of view to verify the field dependence of the grism properties .
the next step is therefore to be able to extract and calibrate a spectrum in wavelength and flux _ at any position in the field of view_. this is achieved by interpolating each of the quantities derived in the previous sections across the chip as a function of the ( x , y ) coordinates of the calibration stars , wr45 and wr96 , in the direct image .
each interpolation describes the smooth spatial variation of the grism properties .
specifically , the quantities that need to be interpolated are : the spectral tilt , the y - offset and , for each order , the dispersion ( the second term of the dispersion when available ) and wavelength zero point . in the case of the wfc , the observed positions allowed us to parameterize the grism field dependence with second - order 2d polymonials , which , in the case of the grism 1@xmath14 order , are characterized by an error of [email protected] in the spectral tilt , 0.10 pixels in the y - offset , 10 and 5 in the wavelength zero point in chip 1 and 2 , respectively , 0.6@xmath10 and 0.4@xmath10 in the first ( @xmath30 ) term of the dispersion in chip 1 and 2 , respectively .
the second ( @xmath31 ) term is characterized by a r.m.s of 27@xmath10 and 22@xmath10 in chip 1 and 2 , respectively ( larsen & walsh 2005 ) .
these interpolations were used to extract the spectra of the planetary nebula smp-81 , which in turn confirmed the consistency of the grism wavelength calibration . since
only 5 positions were acquired across the hrc field of view , we interpolated the grism properties with a first - order 2d polynomial , which , for the 1@xmath14 order , has errors of [email protected] in the spectral tilt , 0.4 pixels in the y - offset , 4.8 ( i.e. 0.2 pixels ) in the wavelength zero point and 0.2@xmath10 and 10@xmath10 in the first ( @xmath30 ) and second ( @xmath31 ) terms of the dispersion ( pasquali et al .
2003b ) .
the coefficients of the 2d polynomials fitting the field dependence of the grism properties are stored in _ configuration files _
( one for each wfc chip and one for hrc ) which are used by axe and can be retrieved at http://www.stecf.org/software/axe .
the configuration files contain three other fundamental keywords : the name of the flat - field cube to be applied for the flat - field correction , the name of the sensitivity function for the flux calibration and the length of the orders in pixels . to test the consistency of our interpolation , we re - extracted the spectra of wr45 and wr96 with axe and the configuration files described above .
figure 14 shows the results for the different pointings across chip 1 of the wfc .
the spectra coincide within 0.4 [ @xmath18 16 , or 0.2 pixels ( @xmath18 5 ) in the case of the hrc ] .
observations with the acs grism require science targets to be imaged through a filter ( direct image ) and the grism ( grism image ) .
the reason is that the target position in the direct image defines the zero point in the wavelength calibration and the flat - field correction for the flux calibration of a spectrum in the grism image .
calibration data have shown that the object position in the direct image is independent of the filter used for imaging and is stable within 25 mas ( 0.5 pixels ) when multiple acquisitions are performed within the same orbit .
direct images taken of the same target with the same postarg values during different orbits are instead stable within 100 mas ( 2 pixels ) .
while a shift of @xmath40 0.5 pixels ( i.e. about 20 in the grism first order spectrum ) is within the uncertainties of the wavelength calibration , an offset of 2 pixels introduces a systematic error of @xmath18 80 in the wavelength calibration of the grism 1@xmath14 order .
it is thus advisable to always acquire a direct image of the targeted field together with its grism images . in the case of multiple exposures of the same field with dither shifts
, it suffices to have one direct image corresponding to one of the grism images and to use the 0@xmath17 orders to confirm the telescope shifts ( e.g. this technique was applied by pirzkal et al .
2004 to the grapes data ) .
direct images can be combined to increase the s / n threshold for the detection of faint objects , but the positions of the objects measured in the final drizzled version have to be related to positions in the raw , geometrically distorted frame for a correct use of axe ( cf .
pirzkal et al .
a spectroscopic drizzle task has been implemented in the latest version of axe ( axe-1.4 , kmmel et al .
2004 ) , which extracts 2d stamp images of a source spectrum appearing in multiple grism images , fully calibrated in wavelength and corrected for flat field .
in the most general case , the source has been imaged in different positions across the field of view , therefore , because of the field dependence of the spectral dispersion , its spectra are characterized by different spectral dispersions .
the axe drizzle corrects the tilt of the individual spectral stamp image , so that its dispersion and the cross - dispersed axes are parallel to the x and y axes , respectively .
it also makes the wavelength scale and the pixel scale in the cross - dispersed direction the same in all the spectral stamp images of the source .
these drizzled stamp images are then coadded , and the final 1d spectrum of the source is extracted and flux calibrated , using a weighting scheme based on the s / n variations as provided by the drizzle weights ( hence proportional to the different exposure times of the observations ) .
users can set the wavelength scale and the pixel scale in the cross - dispersed direction in the axe configuration file ; this option allows to enhance the resolution ( spectral and angular ) of the extracted spectra if the original data were taken with subpixel stepping .
the acs slitless grism is equivalent to a ground - based multi - object spectrograph .
the absence of a slit , though , affects the grism resolution , which the object size and inclination with respect to the dispersion axis can easily decrease ( cf .
this effect can be modeled by , for example , deconvolving the spectrum by the spatial profile of the source , once the latter is known as a function of wavelength .
in addition , the large extension of the grism orders may cause spectra of two adjacent objects to overlap . in its present form axe provides a quantitative estimate of contamination but it does not correct for spectra contamination .
therefore it is suggested to acquire data using different telescope roll angles so to vary the direction of the dispersion axis and reduce the amount of spectra overlap .
also , orienting the telescope so that the dispersion axis is aligned with the object minor axis , certainly maximizes the spectral resolution achievable for that target .
a final note has to be spent on the background level .
since each pixel of the spectrum sees the background emission integrated over the whole spectral range of the grism , the use of a high s / n background image , properly scaled to the exposure time of the data , will improve the s / n in the extracted spectra of faint sources .
as already mentioned before , acs slitless grism spectroscopy has extensively been carried out by large programs such as apples ( acs pure parallel ly@xmath13 emission survey , pi rhoads , i d 9482 ) , grapes ( grism acs program for extragalactic science , pi malhotra , i d 9793 ) and pears ( probing evolution and reionization spectroscopically , pi malhotra , i d 10530 ) .
grapes in particular has proved that the acs grism can detect the continuum emission of sources as faint as z@xmath41 = 27.2 mag and up to @xmath42 7 using 40 orbits ( pirzkal et al .
this is the deepest slitless spectroscopy survey to date , even when compared with ground - based telescopes .
it has allowed to characterize several types of galaxy population : lyman break and ly@xmath13 galaxies at 4 @xmath9 7 ( rhoads et al . 2005 ) , old stellar populations at 0.5 @xmath9 2.5 ( daddi et al . 2005 , pasquali et al . 2005 ) , and emission line galaxies at @xmath43 1.5 ( pirzkal et al .
2005b ) , often at a luminosity level fainter than previously reached .
the success of these surveys clearly indicate that the acs grism spectroscopy provides important insights on galaxy evolution at high redshift .
we would like to thank f. paresce and f. van den bosch for inspiring discussions and comments , and an anonymous referee for improving the paper .
we acknowledge support from f. pierfederici and m. dolenski in querying the st - ecf archive , and help from s. baggett and p. goudfrooj on hst documentation .
acker , a. , ochsenbein , f. , stenholm , b. et al .
, 1992 , strasbourg - eso catalogue of galactic planetary nebulae ( eso ) blakeslee , j.p . ,
tsvetanov , z.i .
, riess , a.g .
et al . , 2003 , apj , 589 , 693 ciardullo , r. , jacoby , g.h . , ford , h.c . , neill , j.d . , 1989 , apj , 339 , 53 daddi , e. , renzini , a. , pirzkal , n. et al . , 2005 , apj , 626 , 680 de mello , d.f . , and pasquali , a. 2001 , a&a , 378 , l10 freudling , w. , 1999 , nicmoslook user manual 3.1 freudling , w. , pirzkal , n. , 1997 , nicmoslook user manual 1.1 fruchter , a.s . ,
hook , r.n .
, 1998 , astro - ph/9808087 hack , w. 1992 , foc isr @xmath21065 hack , w. 1996 , foc isr @xmath21092 hook , r.n . ,
fruchter , a.s .
, 2000 , in astronomical data analysis software and systems ix , asp conference proceedings , 216 , 521 horne , k. , and mackenty , j. 1991 , wfpc1 isr @xmath21008 jacoby , g.h . , ciardullo , r. , 1999 , apj , 515 , 169 jakobsen , p. , albrecht , r. , barbieri , c. et al .
1993 , , 417 , 528 jansen , r.a . , and jakobsen , p. 2001
, a&a , 370 , 1056 kmmel , m. , walsh , j.r . , larsen , s. , hook , r. , 2004 , st - ecf newsletter 36 , 10 larsen , s. , walsh , j.r .
, 2005 , updated wavelength calibration for the wfc / g800l grism , st - ecf isr acs , in preparation magee , d. , bouwens , r. , illingworth , g. et al . , 2002 , iau circ 7908 malhotra , s. , rhoads , j.e . , pirzkal , n. et al . , 2005 , apj , 626 , 666 mccarthy , d.w . , campins , h. , kern , s. et al . , 1998 , aas , dps , 30 , 1114 mccarthy , p.j
. , yan , l. , freudling , w. et al . , 1999 , , 520 , 548 mutchler , m. , cox , c. , 2001 , stsci isr acs 2001 - 07 noll , k. et al . , 2004 , nicmos instrument handbook , version 7.0 , ( baltimore : stsci ) paresce , f. , and greenfield p. 1989 , foc isr @xmath21038 pasquali , a. , ferreras , i. , panagia , n. et al . , 2005b , apj , accepted pasquali , a. , larsen , s. , ferreras , i. et al . , 2005a , aj , 129 , 148 pasquali , a. , pirzkal , n. , walsh , j.r .
et al . , 2001a , st - ecf isr acs 2001 - 02 pasquali , a. , pirzkal , n. , walsh , j.r . , 2001b , st - ecf isr acs 2001 - 04 pasquali , a. , pirzkal , n. , walsh , j.r . , 2003a , st - ecf isr acs 2003 - 01 pasquali , a. , pirzkal , n. , walsh , j.r .
, 2003b , st - ecf isr
acs 2003 - 02 pirzkal , n. , pasquali , a. , 2001 , st - ecf newsletter 28 , p. 3
pirzkal , n. , pasquali , a. , demleitner , m. , 2001a , st - ecf newsletter 29 , p. 5
pirzkal , n. , sahu , k.c . ,
burgasser , a. et al . , 2005a , apj , 622 , 319 pirzkal , n. , pasquali , a. , walsh , j.r . , et al . , 2001b , st - ecf isr
acs 2001 - 01 pirzkal , n. , pasquali , a. , walsh , j.r .
, 2003a , in 2002 hst calibration workshop , stsci 2002 , eds .
s. arribas , a. koekemoer & b. whitmore , p. 74
pirzkal , n. , pasquali , a. , walsh , j.r . , 2003b , st - ecf isr
acs 2003 - 04 pirzkal , n. , xu , c. , ferreras , i. et al . , 2005b , apj , accepted ( astro - ph/0509551 ) pirzkal , n. , xu , c. , malhotra , s. et al . , 2004 , apjs , 154 , 501 plait , p. 1998a
, aas , 192 , 6706 plait , p. 1998b
, aas , 193 , 9802 press , w. h. , teukolsky , s. a. , vetterling , w. t. , & flannery , b. p. , 1992
, numerical recipes in c / fortran , cambridge university press rhoads , j.e . , panagia , n. , windhorst , r.a .
2005 , apj , 621 , 582 riess , a.g . ,
strolger , l .- g . ,
tonry , j. et al . , 2004 ,
apj , 600 , l163 teplitz , h.i . ,
collins , n.r . , gardner , j.p . ,
rhoads , j.e . , 2001 ,
aj , 122 , 1023 teplitz , h.i . ,
collins , n.r . , gardner , j.p .
et al . , 2003 apjs , 146 , 209 tsvetanov , z. , blakeslee , j. , ford , h. et al . , 2002 ,
7912 van der hucht , k.a . , 2001 , the viith catalogue of galactic wolf - rayet stars , new ar , 45 , 135 walsh , j.r . ,
freudling , w. , pirzkal , n. , pasquali , a. , 2003 , st - ecf isr acs 2003 - 03 walsh , j.r . , pirzkal , n. , 2005 , st - ecf isr acs 2005 - 02 windhorst , r. , bernstein , r.a . ,
collins , n. et al . , 2000 , aas , 197 , 12301 xu , c. , rhoads , j.e . , malhotra , s. et al .
, 2005 , apj submitted young , p.r . , and dupree , a.k .
2002 , , 565 , 598 lrrr wfc : & & & + wr45 & 1 s & 1 s & 20 s + wr96 & & 1 s & 20 s + wr96 & & 1 s & 15 s + lmc - smp-81 & & 15 s & 200 s + g191b2b & & 1 s & 15 s + gd153 & & 2 s & 60 s + & & & + hrc : & & & + wr45 & & 3 s & 60 s + gd153 & & 6 s & 180 s + ccccccccc wfc / chip1 & & & & & & & & + ( 29 , 1883 ) & 4762.17 & 35.50 & 0.00414 & 2546.53 & 16.91 & 0.00306 & 1529.99 & 12.46 + ( 29 , 1883 ) & 4763.12 & 35.49 & 0.00432 & 2550.02 & 16.88 & 0.00314 & 1537.05 & 12.44 + ( 54 , 296 ) & 4769.58 & 37.27 & 0.00480 & 2634.18 & 16.91 & 0.00549 & 1537.22 & 13.06 + ( 1939 , 1058 ) & 4733.32 & 39.93 & -0.00007 & 2628.73 & 17.82 & 0.00556 & 1541.92 & 13.64 + ( 1940 , 1057 ) & 4738.62 & 39.98 & -0.00025 & 2636.27 & 17.76 & 0.00570 & 1540.71 & 13.65 + ( 3396 , 1638 ) & 4765.84 & 40.21 & 0.00542 & 2652.97 & 18.11 & 0.00657 & 1552.87 & 14.03 + ( 3396 , 1638 ) & 4764.46 & 40.24 & 0.00519 & 2653.43 & 18.08 & 0.00666 & 1538.95 & 14.07 + ( 3381 , 269 ) & 4776.26 & 41.87 & 0.00771 & 2633.99 & 19.22 & 0.00644 & 1543.18 & 14.71 + ( 2950 , 223 ) & 4778.30 & 41.81 & 0.00414 & 2708.75 & 18.15 & 0.00865 & 1546.97 & 14.53 + ( 3662 , 1820 ) & 4751.05 & 40.61 & 0.00373 & 2577.85 & 18.99 & 0.00437 & - & - + ( 2959 , 1778 ) & 4769.01 & 39.67 & 0.00367 & 2629.52 & 18.08 & 0.00552 & 1565.00 & 13.75 + ( 163 , 1910 ) & 4756.89 & 35.98 & 0.00279 & 2508.35 & 17.42 & 0.00202 & 1564.33 & 12.44 + ( 1111 , 1915 ) & 4768.77 & 36.99 & 0.00420 & 2607.01 & 17.06 & 0.00446 & 1575.82 & 12.83 + ( 1125 , 337 ) & 4776.17 & 38.97 & 0.00352 & 2657.27 & 17.41 & 0.00635 & 1553.18 & 13.55 + ( 2044 , 257 ) & 4769.19 & 40.56 & 0.00302 & 2710.59 & 17.52 & 0.00817 & 1547.22 & 14.05 + ( 2045 , 1054 ) & 4774.98 & 39.24 & 0.00451 & 2665.68 & 17.50 & 0.00664 & 1559.31 & 13.66 + wfc / chip2 & & & & & & & & + ( 3332 , 1784 ) & 4746.52 & 43.32 & 0.00334 & 2699.07 & 18.72 & 0.00917 & 1548.27 & 14.94 + ( 3322 , 400 ) & 4777.15 & 44.73 & 0.00806 & 2728.65 & 19.34 & 0.01084 & 1536.41 & 15.71 + ( 3322 , 400 ) & 4782.92 & 44.67 & 0.00852 & 2727.76 & 19.36 & 0.01081 & 1544.99 & 15.69 + ( 303 , 1779 ) & 4761.05 & 38.56 & 0.00327 & 2609.03 & 17.63 & 0.00511 & 1389.09 & 13.77 + ( 314 , 380 ) & 4782.26 & 39.77 & 0.00721 & 2614.51 & 18.40 & 0.00555 & 1532.86 & 14.04 + ( 314 , 380 ) & 4772.47 & 39.86 & 0.00669 & 2665.29 & 17.79 & 0.00720 & 1553.76 & 13.98 + ( 2023 , 1934 ) & 4774.60 & 40.78 & 0.00462 & 2701.40 & 17.79 & 0.00816 & 1551.22 & 14.20 + ( 2024 , 1934 ) & 4775.98 & 40.94 & 0.00348 & 2698.68 & 17.85 & 0.00805 & 1563.61 & 14.18 + ( 3642 , 309 ) & 4758.08 & 45.90 & 0.00491 & 2730.46 & 19.61 & 0.01124 & 1541.64 & 15.93 + ( 2954 , 253 ) & 4774.10 & 44.81 & 0.00476 & 2764.58 & 18.74 & 0.01207 & 1555.78 & 15.54 + ( 225 , 289 ) & 4775.20 & 39.98 & 0.00561 & 2607.48 & 18.45 & 0.00548 & 1529.12 & 14.04 + ( 1151 , 327 ) & 4786.96 & 41.63 & 0.00497 & 2646.24 & 18.83 & 0.00664 & 1534.97 & 14.55 + ( 2059 , 268 ) & 4788.72 & 43.10 & 0.00545 & 2672.69 & 19.25 & 0.00793 &
1559.76 & 15.02 + ccccccc wfc / chip1 & & & & & & + ( 29 , 1883 ) & - & - & - & - & - & - + ( 29 , 1883 ) & - & - & - & - & - & - + ( 54 , 296 ) & - & - & - & - & - & - + ( 1939 , 1058 ) & -4965.81 & -40.61 & -2317.55 & -19.66 & -1472.93 & -12.91 + ( 1940 , 1057 ) & -5045.09 & -40.83 & -2319.40 & -19.65 & -1472.69 & -12.90 + ( 3396 , 1638 ) & -5069.28 & -42.43 & -2355.43 & -20.47 & -1531.79 & -13.49 + ( 3396 , 1638 ) & -5081.47 & -42.47 & -2374.06 & -20.52 & -1565.35 & -13.54 + ( 3381 , 269 ) & -5123.16 & -44.59 & -2340.16 & -21.39 & -1530.69 & -14.13 + ( 2950 , 223 ) & -5115.70 & -43.85 & -2376.59 & -21.14 & -1485.96 & -13.84 + ( 3662 , 1820 ) & -5094.49 & -42.70 & -2377.55 & -20.61 & -1490.10 & -13.50 + ( 2959 , 1778 ) & -5077.32 & -41.56 & -2359.10 & -20.06 & -1554.45 & -13.26 + ( 163 , 1910 ) & - & - & - & - & - & - + ( 1111 , 1915 ) & -5046.21 & -38.53 & -2334.57 & -18.56 & -1287.35 & -11.90 + ( 1125 , 337 ) & -5048.17 & -40.47 & -2348.77 & -19.53 & -1565.10 & -12.91 + ( 2044 , 257 ) & -5110.14 & -42.33 & -2367.31 & -20.38 & -1494.76 & -13.36 + ( 2045 , 1054 ) & -5091.67 & -41.17 & -2359.96 & -19.84 & -1494.91 & -13.00 + wfc / chip2 & & & & & & + ( 3332 , 1784 ) & -5094.14 & -45.24 & -2375.18 & -21.84 & -1623.83 & -14.51 + ( 3322 , 400 ) & -5129.59 & -47.57 & -2377.74 & -22.90 & -1666.43 & -15.31 + ( 3322 , 400 ) & -5096.84 & -47.42 & -2381.09 & -22.91 & -1621.90 & -15.23 + ( 303 , 1779 ) & - & - & - & - & - & - + ( 314 , 380 ) & - & - & - & - & - & - + ( 314 , 380 ) & - & - & - & - & - & - + ( 2023 , 1934 ) & -5059.91 & -42.64 & -2352.79 & -20.57 & -1460.89 & -13.46 + ( 2024 , 1934 ) & -5069.51 & -42.64 & -2356.22 & -20.57 & -1478.50 & -13.48 + ( 3642 , 309 ) & -5310.33 & -49.11 & -2381.69 & -23.31 & -1482.93 & -15.25 + ( 2954 , 253 ) & -5111.24 & -47.01 & -2548.88 & -23.09 & -1456.04 & -14.79 + ( 225 , 289 ) & - & - & - & - & - & - + ( 1151 , 327 ) & -5056.06 & -43.33 & -2352.71 & -20.91 & -1522.75 & -13.77 + ( 2059 , 268 ) & -5109.74 & -45.28 & -2369.10 & -21.81 & -1463.58 & -14.24 + cccccc ( 572 , 574 ) & 4755.91 & 23.86 & 0.0018 & 2475.29 & 12.00 + ( 207 , 944 ) & 4749.06 & 24.59 & 0.0019 & 2471.58 & 12.35 + ( 754 , 391 ) & 4759.68 & 23.51 & 0.0019 & 2480.18 & 11.82 + ( 199 , 391 ) & 4756.92 & 23.87 & 0.0018 & 2476.84 & 11.99 + ( 764 , 949 ) & 4753.80 & 24.19 & 0.0021 & 2471.41 & 12.19 + | the advanced camera for surveys on - board _ hst _ is equipped with a set of one grism and three prisms for low - resolution , slitless spectroscopy in the range 1150 to 10500 .
the g800l grism provides optical spectroscopy between 5500 and @xmath0 1 @xmath1 m with a mean dispersion of 39 / pix and 24 / pix ( in the first spectral order ) when coupled with the wide field and the high resolution channels , respectively .
given the lack of any on - board calibration lamps for wavelength and narrow band flat - fielding , the g800l grism can only be calibrated using astronomical targets . in this paper
, we describe the strategy used to calibrate the grism in orbit , with special attention to the treatment of the field dependence of the grism flat - field , wavelength solution and sensitivity in both channels . |
the production of final state photons at large transverse momenta in high - energy processes is an important observable for testing qcd .
data on large-@xmath3 photon production in hadronic collisions have been used in the past to obtain information on the gluon distribution in the photon .
in addition , a good understanding of direct photon production within the framework of the standard model will be essential for new physics searches at future colliders .
photons from high - energy collisions are produced essentially by two mechanisms : the direct production of a photon off a primary quark or antiquark , or the fragmentation of a hadronic jet into a single photon carrying a large fraction of the jet energy .
the direct production gives rise to perturbatively calculable short - distance contributions whereas the fragmentation contribution is due primarily to a long distance process which can not be calculated completely inside perturbative qcd .
the latter is described by the process - independent quark- , antiquark- , or gluon - to - photon fragmentation functions ( ff ) @xcite which must be fixed by experimental data as long as they can not be calculated by non - perturbative methods . their evolution with the factorization scale @xmath4 and their so - called point - like contribution ( up to a normalization )
can , however , be calculated perturbatively .
directly produced photons are usually well separated from hadron jets , while photons originating from the fragmentation process are primarily found inside hadronic jets . by imposing an isolation criterion on the photon
, one can , in principle , suppress ( but in general , not eliminate ) the fragmentation contribution to the final state photon cross section .
so far , only a limited number of measurements of single photon production exists through which direct information on the photon ff can be obtained .
a possible way is the measurement of the inclusive photon cross section in @xmath5 annihilation .
recently the opal collaboration at lep has measured this cross section for final state photons in the range @xmath6 , where @xmath7 is the photon energy fraction in terms of the beam energy @xcite .
the results are in reasonable agreement with predictions obtained within perturbative qcd using the model - dependent parametrizations of the photon ff of do @xcite , grv @xcite or bfg @xcite , when choosing the factorization scale @xmath8 .
the do model for the ff is based on an asymptotic solution of the evolution equation whereas the other two models @xcite contain in addition a non - perturbative input inspired by the vector meson dominance model .
unfortunately , the experimental precision was not sufficiently high to discriminate between these theoretical predictions .
note also that in @xcite it was shown that predictions obtained within perturbative qcd and using a fixed - order expanded expression for the photon ff also agreed well with the opal data .
an alternative way to determine the ff from @xmath5 data is to measure the production of photons accompanied with a definite number of hadronic jets @xcite . in this approach
, the photon is treated like any other hadron and is clustered simultaneously with the other hadrons into jets ( so - called democratic clustering procedure ) @xcite .
then , one of the jets in the final state is considered as the photon jet if the fraction of the electromagnetic energy inside the jet is sufficiently large , i.e. @xmath9 with @xmath10 fixed by the experimental conditions . on this basis the aleph collaboration at lep analysed events produced at the @xmath11-resonance @xcite which contained one hadron jet and a jet with a photon carrying at least @xmath12 of the jet energy .
this @xmath13-jet rate was used to determine the quark - to - photon ff , which was calculated in leading order ( lo ) , i.e. up to o(@xmath14 ) , in @xcite .
the calculation was extended to next - to - leading order ( nlo ) , i.e. up to o(@xmath15 ) , in @xcite and a nlo photon - to - quark ff was obtained @xcite by comparing with the aleph data . due to the unfavorable signal - to - background ratio below @xmath16 for this particular observable , the @xmath13-jet rate , the photon ff could be determined only in the range @xmath17 . in @xcite also
the predictions of grv @xcite and bfg @xcite for the photon ff were compared to the aleph @xmath13-jet data .
it turned out that the bfg parametrization is in agreement with the measurements while the rate predicted with the help of the ff of grv lies too high .
the photon ff is process independent as any other ff and can be used to predict the cross section for the production of single photons in other processes .
this has been done recently for isolated photon production in deep - inelastic @xmath18 scattering ( dis ) @xcite . in that
work the emphasis was put on the prediction of cross sections in the hera kinematic range without the necessity to introduce photon - parton separation cuts as used in earlier work @xcite .
this was achieved by including the divergent photon - quark ( -antiquark ) singular contributions and the bare quark - to - photon fragmentation contributions , leaving in the final result only a finite factorization scale dependent quark - to - photon ff contribution . in our previous work we adopted as the photon
ff the lowest - order fit obtained from the photon+1-jet data of the aleph collaboration @xcite . in the present work we study the influence of the fragmentation contribution on the cross section for the production of a photon plus jets in dis and give predictions for @xmath19 as a function of @xmath1 in the interval @xmath20 ( for the definition of @xmath1 see below ) . for this purpose
we calculate @xmath19 with the same cuts on the dis variables as in @xcite for essentially three different sets of photon ff s , the aleph sets in lo and nlo @xcite , the leading - logarithmic ( ll ) and beyond - leading logarithmic ( bll ) parametrizations of grv @xcite and the bll parametrization of bfg @xcite .
the first set of photon ff is obtained within a fixed - order framework at a given order in @xmath21 .
the two other sets were obtained after the leading and/or next - to - leading logarithms of the factorization scale @xmath22 were resummed .
the essence of these two approaches has been described extensively in @xcite and will be outlined below .
in leading order from the @xmath23 vertex . ]
( i.e. , at o(@xmath14 ) ) , the cross section for the production of photon plus jets in dis receives contributions from the quark ( antiquark ) subprocess @xmath24 ( @xmath25 ) .
together with the remnant jet from the proton this process gives rise to a @xmath26-jet final state . in the virtual photon - proton center - of - mass system
the hard @xmath27 recoils against the hard jet back - to - back .
cuts on the usual dis variables @xmath28 , @xmath29 and @xmath0 are applied to remove @xmath27 production by incoming @xmath30 s of small virtuality . to produce @xmath27 s of sufficiently high energy , an explicit cut on the total @xmath31 center - of - mass energy , @xmath32 ,
is introduced .
both leptons and quarks emit photons .
the leptonic radiation is suppressed by a cut on the photon emission angle with respect to the incoming electron in the same way as in the earlier work @xcite . in lo
each parton is identified with a jet and the photon is automatically isolated from the quark jet by demanding a non - zero transverse momentum of the photon or jet in the @xmath31 center - of - mass frame . at this order
there is no fragmentation contribution . at next - to - leading order ( o(@xmath15 ) ) we have subprocesses with an additional gluon , either in the final or in the initial state , i.e. the subprocesses @xmath33 ( and similarly with @xmath34 replaced by @xmath35 ) and @xmath36 , respectively . in addition , virtual corrections ( one - loop diagrams at o(@xmath15 ) ) to the lo processes have to be included .
the processes @xmath33 and @xmath36 contribute both to the @xmath26-jets cross section and to the cross section for @xmath37-jets . in the latter case each parton in the final state
( including the photon ) builds a jet on its own , whereas for @xmath26-jets a pair of final state partons is combined into one jet .
the recombination of two partons will be performed with the help of the cone algorithm .
the exact prescriptions will be given later when we present our results .
in the nlo calculation of the @xmath26-jet cross section one encounters the well known infrared and collinear singularities .
for the processes @xmath33 and @xmath36 they appear in those phase space regions where two partons are degenerate to one parton , i.e. when one of the partons becomes soft or two partons become collinear to each other .
these singularities cancel in the case of soft gluons or for collinear quark - gluon pairs against the singularities from the virtual corrections to the lo process or have to be factorized and absorbed into the renormalized parton distribution functions ( pdf s ) of the proton . to accomplish this cancellation ,
the singularities are isolated in an analytic calculation with the help of dimensional regularization and the phase space slicing method .
the technical details of this procedure have been described in @xcite and need not to be repeated here .
the matrix elements @xmath38 for the processes @xmath33 and @xmath36 have also photonic infrared and collinear singularities .
the infrared singularity is outside the kinematical region we are interested in since the photon is required to be observed in the detector . in the numerical calculation of the cross section
we impose this condition by demanding a minimal transverse momentum of the final state photon .
this cut removes also collinear singularities due to initial state photon radiation .
final state collinear singularities are present and are treated with the phase space slicing method similar to the case of the quark - gluon collinear contribution . for this purpose the phase space slicing parameter @xmath39 is introduced , which is chosen very small , so that the matrix element @xmath38 can be approximated by its singular part .
the slicing of the phase space is done with the help of the scaled squared invariant masses , for example , with @xmath40 for the subprocess @xmath33 , where @xmath41 and @xmath42 are the four - momenta of the outgoing photon and quark , respectively . in
the gluon initiated subprocess one has two singular regions , which are controlled by the variables @xmath43 and @xmath44 , respectively . in the region @xmath45
the cross section is evaluated numerically .
the details for the calculation of the various contributions are described in @xcite .
the contribution to @xmath38 in the region @xmath46 is collinearly divergent and is regulated by dimensional regularization .
the divergent part is absorbed into the bare photon ff to yield the renormalized photon ff denoted by @xmath47 .
the additional fragmentation contribution to @xmath38 for the subprocess @xmath33 has the following form @xmath48 there is a similar expression for the subprocess @xmath36 .
it is obvious from ( [ m2frag ] ) that the fragmentation contribution is o(@xmath49 ) , the photon ff @xmath50 given by @xmath51 being of o(@xmath14 ) .
@xmath52 in ( [ dqg ] ) stands for the non - perturbative ff of the transition @xmath53 at the factorization scale @xmath22 , i.e. the scale at which the redefinition has been performed .
this function will be given by one of the sets of photon ff mentioned above . the second term in ( [ dqg ] ) ,
if substituted in ( [ m2frag ] ) , is the finite part resulting from adding the bare photon ff and the collinear photon - quark ( -antiquark ) contribution to the matrix element @xmath54 integrated in the region @xmath46 . the @xmath39 dependence in ( [ dqg ] ) is expected to cancel the dependence of the numerically evaluated @xmath26-jet cross section restricted to the region @xmath55 as studied in @xcite .
this means that @xmath39 is only a technical cut separating divergent from finite contributions .
this collinear photon - quark ( -antiquark ) contribution to the matrix element @xmath56 being calculated in the collinear approximation is valid only up to terms o(@xmath39 ) .
consequently the cut @xmath39 must be chosen sufficiently small .
the @xmath39 independence must be checked by varying @xmath39 .
results of this test will be shown below . in ( [ dqg ] )
, @xmath57 is the lo quark - to - photon splitting function @xmath58 and @xmath59 is the electric charge of quark @xmath34 .
the variable @xmath60 denotes the fraction of the final photon energy in terms of the energy of the quark emitting the photon .
if the photon is emitted from a final state quark with four - momentum @xmath61 , then @xmath60 is obtained from @xmath62 .
it can also be related to the invariants @xmath63 and @xmath64 obtained from the four - momenta of the subprocess @xmath65 with the result @xmath66 according to ( [ m2frag ] ) the fragmentation contribution to the subprocess @xmath67 is calculated from the convolution of the ff with the o(@xmath68 ) matrix element @xmath38 of the process @xmath69 , which is well known .
it yields a term of o(@xmath49 ) to the cross section for @xmath26-jets .
similarly the fragmentation contribution to the subprocess @xmath36 is calculated from the convolution of the ff for @xmath70 with the matrix element @xmath38 of the process @xmath71 which is also known .
equivalent formulas are used for the calculation of the fragmentation contribution to @xmath72 and @xmath36 , where the @xmath35 fragments into a photon .
the fragmentation is always assumed to occur collinearly , i.e. with no additional transverse momentum in the fragmentation process .
as long as we intend to calculate the @xmath73-jet cross section only up to o(@xmath49 ) we need the photon ff only in lo , i.e. at o(@xmath14 ) .
this means that the gluon - to - photon ff , which starts at o(@xmath49 ) , is not needed . in
lo the non - perturbative quark - to - photon ff obeys the evolution equation @xmath74 with the solution @xmath75 @xmath76 is an initial value which must be fitted to experimental data with a chosen initial scale @xmath77 .
this has been done by the aleph collaboration @xcite . from the fit to their @xmath78-jet data @xcite they obtained @xmath79 with @xmath80 gev .
( [ dqg - solu ] ) together with ( [ dqg - lo ] ) is one of the photon ff choices which will be used to predict the cross section @xmath19 for @xmath73-jet production in dis @xmath18 scattering .
we note that ( [ dqg - solu ] ) is an exact solution of ( [ dqg - evol ] ) at o(@xmath14 ) .
furthermore when we substitute ( [ dqg - solu ] ) into ( [ m2frag ] ) we see that together with the finite contribution ( i.e. , the second term in ( [ dqg ] ) ) the cross section becomes independent of the factorization scale @xmath4 .
this means that for the cancellation of the @xmath4 dependence only the lo ff is needed . nonetheless , in order to see the influence of the nlo corrections to @xmath81 we shall evaluate the @xmath73-jet cross section also with the inclusion of the nlo photon ff . similarly to ( [ dqg - solu ] )
the nlo ff @xmath82 is obtained as the solution of ( [ dqg - evol ] ) , but now with o(@xmath49 ) terms added on the right - hand side of ( [ dqg - evol ] ) .
the result at scale @xmath4 is @xmath83 \ln\left(\frac{\mu_f^2 } { \mu_0 ^ 2}\right ) \\
\displaystyle + \frac{\alpha_s}{2\pi } c_f p^{(0)}_{qq}(z ) \ln\left(\frac{\mu_f^2}{\mu_0 ^ 2}\right ) \otimes \left[\frac{\alpha e_q^2}{2\pi } \frac{1}{2}p^{(0)}_{q\gamma}(z ) \ln \left(\frac{\mu_f^2}{\mu_0 ^ 2}\right ) + d_{q \rightarrow \gamma}(z,\mu_0 ) \right ] + d_{q \rightarrow \gamma}(z,\mu_0 ) \end{array } \label{dqg - solu2}\ ] ] where @xmath84 is the next - to - leading order quark - to - photon splitting function @xcite and @xmath85 is the well - known lo @xmath86 splitting function . @xmath87
is the initial value of the nlo ff , which contains all unknown long - distance contributions .
the result in ( [ dqg - solu2 ] ) is an exact solution of the evolution equation up to o(@xmath88 ) .
based on the above arguments , also the nlo photon ff has recently been determined @xcite using the aleph @xmath89-jet data @xcite . a three parameter fit with @xmath90 ( this value for @xmath68 , just a scale - independent parameter here ,
was chosen so that the observed total @xmath5 annihilation cross section into hadrons is reproduced in the o(@xmath68 ) calculation ) yields @xcite @xmath91 with @xmath92 gev . inside the experimental errors
this fit for the photon ff at @xmath93 describes the aleph data as good as the lo fit ( [ dqg - lo ] ) @xcite .
a fit with a larger value of @xmath68 is also reported in @xcite .
it should be noted that the lo and nlo ff s of the photon as given in ( [ dqg - solu ] ) together with ( [ dqg - lo ] ) and in ( [ dqg - solu2 ] ) together with ( [ dqg - nlo ] ) do not take into account resummation of powers of @xmath94 as usually implemented , e.g. via altarelli - parisi evolution @xcite .
such resummations are only unambiguous if the resummed logarithm is the only large logarithm in the kinematical region under consideration .
if logarithms of different arguments can become simultaneously large , the resummation of one of these logarithms at a given order implies that all other potentially large logarithms are shifted into a higher order of the perturbative expansion , i.e. are neglected . in the evaluation of the @xmath89-jet rate at @xmath95 @xcite and @xmath96 @xcite at lep for @xmath97 ,
one encounters at least two different potentially large logarithms , @xmath98 and @xmath99 . in the high @xmath60 region , the region
where the photon is isolated or almost isolated it is by far not clear that @xmath98 is the largest logarithm .
choosing not to resum the logarithms of @xmath100 is therefore equally justified for the case of large @xmath101 . in the conventional approach ,
the parton - to - photon ff s @xmath102 satisfy a system of inhomogeneous evolution equations @xcite .
usually these equations are diagonalized in terms of the singlet and non - singlet quark ff s as well as the gluon ff . for the case
that the gluon - to - photon fragmentation is neglected , which we shall do in our application to dis @xmath27+jets production , these equations can be simplified @xcite .
then the flavor singlet and non - singlet quark - to - photon ff s become equal to a unique function @xmath103 .
this function satisfies the all - orders evolution equation @xmath104 which has a similar form as the next - to - leading order evolution in the fixed - order approach , but the coupling @xmath68 is not fixed and now is taken as a function of the factorization scale @xmath4 .
the full solution @xmath105 of the all - orders inhomogeneous evolution equation is a sum of two contributions .
the first term is the point - like ( or perturbative ) part @xmath106 , which in the leading logarithmic ( ll ) approximation is a solution of the inhomogeneous equation ( [ dqg - evol - in ] ) .
the second term is the hadronic ( or non - perturbative ) part @xmath107 , which is the solution of the corresponding homogeneous equation .
this term must be fitted to experimental data to obtain the input at a starting scale . at ll only terms of the form @xmath108 are kept while beyond leading logarithms ( bll ) both the leading and the sub - leading ( @xmath109 ) logarithms of @xmath4 are resummed to all orders in the strong coupling constant @xmath68 .
thus the lo solution ( [ dqg - solu ] ) is the @xmath110 term of the ll approximation , whereas the nlo solution ( [ dqg - solu2 ] ) is the bll approximation up to the order @xmath111 . by dropping the @xmath112 terms in ( [ dqg - solu2 ] )
one can also infer the ll approximation up to the order @xmath111 . in this conventional approach , the quark - to - photon fragmentation function
is regarded as being of @xmath113 .
in fact , the @xmath110 term of the ll approximation is proportional to @xmath114 . at large scales @xmath115 and assuming that @xmath4 can be identified with the scale that determines @xmath68 , i.e. @xmath116 , the ll ff becomes of the order of o(@xmath117 ) .
this motivates the usual statement that the total fragmentation contribution is like o(@xmath14 ) , i.e. of the same order as the lo direct photon contribution .
of course , this statement depends on the considered kinematic region . with this argument
however , one usually justifies the inclusion of terms in the fragmentation contribution of o(@xmath118 ) which become of o(@xmath49 ) if the ff is considered of o(@xmath117 ) .
the most recent parametrizations of photon ff s in the conventional approach as described in the previous paragraph are those of grv @xcite and bfg @xcite .
the ff s of grv @xcite exist in the ll as well as in the bll form , which in the following we shall denote as the lo and nlo ff s of grv .
however , the ff of bfg @xcite are available only in the bll approximation .
they both have been compared to the aleph @xmath119-jet cross section @xcite , which is sensitive to the large @xmath60 region ( @xmath120 ) .
the bfg prediction goes through the experimental points whereas the grv nlo parametrization lies systematically higher and agrees less well with the aleph data @xcite .
this difference can be attributed to both the choice of the input scale @xmath77 and of the non - perturbative input at this scale .
the bfg input is smaller and according to the aleph data this choice is preferred .
thus we have at least five different versions of photon ff s at our disposal which have been compared to the aleph data : two lo parametrizations , the one using the aleph data directly to determine the initial distribution given in ( [ dqg - solu ] ) and ( [ dqg - lo ] ) and the grv parametrization in lo ; and three nlo choices , the one written in ( [ dqg - solu2 ] ) and ( [ dqg - nlo ] ) directly fitted to the aleph data , and the nlo parametrizations of grv and bfg .
we shall use these five parametrizations to predict the @xmath73-jet cross section in deep - inelastic @xmath18 scattering .
the results for these cross sections are obtained for energies and kinematical cuts appropriate for the hera experiments .
the energies of the incoming electron and proton are @xmath121 gev and @xmath122 gev , respectively .
the cuts on the dis variables are as in our previous work @xcite @xmath123 to reduce the background from lepton radiation @xcite we require @xmath124 where @xmath125 is the emission angle of the photon measured with respect to the momentum of the incoming electron in the hera laboratory frame .
the cut on @xmath126 , the angle between the photon and the outgoing electron momentum , suppresses radiation from the final - state electron .
the @xmath127 and the hadron jet @xmath128 are required to have minimal transverse momenta @xmath129 different values of minimal @xmath3 s for the photon and the jet have to be chosen in order to avoid the otherwise present infrared sensitivity of the nlo predictions @xcite .
the pdf s of the proton are taken from mrst @xcite .
@xmath68 is calculated from the two - loop formula with the same @xmath130 value ( @xmath131 mev ) as used in the mrst parametrization .
the scale in @xmath68 and the factorization scales @xmath4 for the proton pdf and the photon ff are equal and fixed to @xmath132 .
we are interested in the differential cross section @xmath19 at nlo ( up to o(@xmath49 ) ) as a function of @xmath1 , where @xmath133 this definition of @xmath1 agrees with the one in ( [ zgamma ] ) for the fragmentation contribution . in ( [ zgamma - pt ] ) , @xmath134 is the transverse momentum of the parton producing hadrons , which is recombined with the photon into the photon jet . here
the photon is treated like any other parton during recombination , so that one of the recombined jets may be the photon jet .
for the recombination of the two partons into a hadron jet or a parton and the photon into the photon jet , we use the cone algorithm of the snowmass convention @xcite .
the recombination is applied in the @xmath135 parton sample and yields a contribution to the @xmath26-jet class .
the @xmath37-jet class consists of the unrecombined contributions of the @xmath135 parton sample .
the recombination is performed in the @xmath31 center - of - mass frame .
two partons @xmath136 and @xmath137 are combined into a jet @xmath128 if they obey the cone constrains @xmath138 and @xmath139 , where @xmath140 @xmath141 and @xmath142 are the rapidity and the azimuthal angle of the recombined jet .
if for example , parton @xmath136 is the @xmath127 , then @xmath128 is the photon jet .
the jet variables @xmath141 and @xmath142 are obtained from the averages of the corresponding variables of the recombined partons @xmath136 and @xmath137 multiplied with their respective @xmath3 values and @xmath143 .
we choose @xmath144 .
the azimuthal angle is defined with respect to the scattering plane defined by the momenta of the ingoing and outgoing electron .
it is known that the cone algorithm is ambiguous for final states with more than three particles or partons .
since we have maximally three partons in the final state this is not relevant in our case .
furthermore it will be no problem to repeat the calculation for any other cluster algorithm that might be used in the analysis of forthcoming experimental data .
it is obvious that for the fragmentation contribution the cone constraint is always satisfied since the hadronic remnant in the fragmentation is collinear with the photon .
( 160,100 ) ( 30,-1 ) before we present our results for @xmath19 for the various choices of the photon ff we show that the cross section is independent of the slicing cut @xmath39 . for this purpose
we have calculated the sum of the cross sections for the production of @xmath73-jets @xmath145 as a function of @xmath39 in the range @xmath146 .
the result is shown in fig.[fig1 ] .
here we have plotted the cross section for the direct contribution with @xmath147 which increases with decreasing @xmath39 and the direct contribution for @xmath148 together with the second term in contribution ( [ dqg ] ) inserted in ( [ m2frag ] ) which we denote by frag(sing) in the figure .
this latter contribution is negative and decreases with decreasing @xmath39 in such a way that the sum of the increasing and decreasing contributions is constant inside the numerical accuracy for @xmath149 .
the sum shown in fig . [ fig1 ] contains also the fragmentation contribution ( [ m2frag ] ) with the photon ff @xmath52 as given in ( [ dqg - solu ] ) and ( [ dqg - lo ] ) ( denoted by frag(fin) in the figure ) .
of course , this contribution is independent of @xmath39 .
it amounts to approximately @xmath150 of the total @xmath73-jet cross section .
the cross sections shown in fig .
[ fig1 ] are calculated with the kinematical cuts specified above and the additional cut @xmath151 . from fig .
[ fig1 ] we conclude that the slicing cut for the photon should be chosen smaller than @xmath152 to obtain a reliable cross section .
for the following results we fixed it to @xmath153 as in our previous work @xcite .
( 160,100 ) ( 30,-1 ) the results for @xmath19 as a function of @xmath1 are shown in figs .
[ fig2 ] , [ fig3 ] and [ fig4 ] .
the total cross section is dominated by the last bin @xmath154 which is clipped off above 10 pb in these figures . in this bin
we have the direct contribution and those fragmentation contributions with no or little hadronic remnant .
the fraction of the total cross section ( for @xmath155 ) contained in the last bin amounts to @xmath156 to @xmath157 and does not depend on the parametrization of the ff .
therefore it is clear that the dependence of the photon plus jet cross section on @xmath1 below @xmath158 is an appropriate observable which contains information on the photon ff . here
the fragmentation contribution dominates and in addition we have those direct contributions from the ( @xmath159)-parton sample where the photon is recombined with one of the partons . in fig .
[ fig2 ] the cross section @xmath19 in the region @xmath160 is plotted for the photon ff s from ( [ dqg - solu ] ) and ( [ dqg - lo ] ) ( denoted aleph ) and from ( [ dqg - solu2 ] ) and ( [ dqg - nlo ] ) ( denoted ggdr(@xmath90 ) in the figure ) .
for both cases the cross section first decreases with increasing @xmath1 until it reaches a minimum near @xmath161 , from which it increases strongly towards @xmath158 .
this dependence is similar to what has been found for the photon plus one - jet cross section in @xmath5 annihilation at the @xmath11-resonance @xcite .
note however , that below @xmath162 , i.e. outside the range shown in these figures , the cross section decreases again towards smaller values of @xmath1 . unlike at lep ,
the photon - jet is here required to have a minimal transverse momentum as given by ( [ zgamma - pt ] ) .
the decrease in the cross section for values of @xmath1 below 0.5 is a kinematical effect resulting from the imposition of this cut .
for the two choices of ff s the predicted cross section differs by approximately @xmath163 .
the cross section for the lo ff is smaller than for the nlo ff .
we note that the @xmath68 value quoted for this ff is not the value for which the cross section was calculated ; rather it is the value which was chosen to fit the photon ff to the aleph data .
both photon ff s fit the aleph data in lo and nlo approximation . in our case
only the lo ff is appropriate .
the result for the nlo ff gives just an indication of higher - order corrections up to o(@xmath118 ) .
however , it is not a full prediction at this order since the nnlo calculation for the direct contribution has not been done , the nlo corrections to @xmath38 on the right - hand - side of ( [ m2frag ] ) are not included either .
( 160,100 ) ( 30,-1 ) in fig .
[ fig3 ] the predictions with grv @xcite and bfg @xcite ff s are shown .
both are ff s in nlo and with all leading and subleading logarithms of @xmath164 resummed at the next - to - leading logarithmic accuracy .
the behavior of the cross section as a function of @xmath1 is qualitatively similar to the result in fig.[fig2 ] .
the cross section for grv is larger than the one with the bfg ff .
this is to be expected since grv gives a larger @xmath5 cross section for @xmath165-jet than bfg also for the aleph kinematical conditions . comparing with the fixed - order result in fig .
[ fig2 ] the bfg cross section lies above the fixed - order prediction obtained using the lo ff over the whole @xmath1 region .
the discrepancy is largest in the region @xmath166 where the cross section is smaller . at the minimum
the two cross sections differ by more than a factor two .
( 160,100 ) ( 30,-1 ) in fig .
[ fig4 ] the results for the grv photon ff s are displayed , but now for both the lo and the nlo parametrization .
on the average the cross section for the lo fragmentation function is again smaller by @xmath167 compared to the nlo prediction .
in addition , we have plotted the cross sections for the point - like approximations of the ff s .
this approximation is rather good , i.e. the influence of the hadronic part of the ff is small in the considered @xmath1 range .
we note that the difference between the full ff and the point - like part increases with decreasing @xmath1 .
if we compare our results for the various photon ff s in figs.[fig2 ] , [ fig3 ] and [ fig4 ] we observe that the predictions agree approximately within @xmath167 in the small and large @xmath1 regions , i.e. for @xmath168 and @xmath169 .
however , near the minimum of the cross section , i.e. in the region @xmath166 , the results differ by up to a factor two .
the largest differences occur between the predictions obtained with the aleph photon fragmentation function on the one side and the grv and bfg parametrizations on the other side .
this difference comes mainly from the fact that different evolution approaches are used .
whereas for grv and bfg the ff at @xmath170 is obtained from the conventional evolution after the leading and/or subleading logarithms of @xmath22 were resummed , the aleph ff s are evolved only to the respective finite order in @xmath21 as given in ( [ dqg - solu ] ) and ( [ dqg - solu2 ] ) . therefore , if we calculated @xmath19 for @xmath171-jets at the large scale @xmath8 the cross sections obtained for bfg and aleph would come out quite similar over the whole @xmath1 range inside the @xmath167 margin .
only when we go to the scale @xmath170 , which on the average is much smaller , we observe that the cross section obtained using the bfg photon fragmentation function is much larger than the aleph cross section in the region @xmath172 . provided the resummed solution of the all - orders evolution equation
can be accurately determined @xcite over the whole @xmath1 range under consideration , i.e. for @xmath173 , the approach using this solution represents the theoretically preferred approach as it is the most complete .
the approach using an expanded and therefore approximated photon ff has however important advantages .
as already mentioned , its use leads to factorization scale independent results for the cross section evaluated at a given fixed order in @xmath21 .
moreover it enables an analytic determination of the photon ff . within the conventional approaches ,
the bfg prediction should be preferred over the result with the grv ff .
the difference between grv and bfg in fig .
[ fig3 ] is related to the choice of a different input at the starting scale @xmath77 .
as already mentioned , the bfg parametrization is in better agreement with the aleph data @xcite than the grv nlo parametrization and therefore is the more realistic choice of ff for the photon .
this could be further tested as soon as data from hera on photon plus jet production in dis become available .
we have presented a nlo calculation for the production of photons accompanied by jets in deep inelastic electron proton scattering , taking into account the contribution from quark - to - photon fragmentation .
we have calculated the cross section as a function of @xmath1 for those parametrizations of the photon fragmentation functions , which had been compared to photon plus jet data from the aleph collaboration . as we observed significant differences between the predictions
, we expect that the measurement of photon plus jet production in dis at hera will contribute to testing these photon fragmentation functions .
j. e. huth et al .
_ summer study on high energy physics , research directions for the decade _ ,
snowmass , 1990 ; f. abe et al . ,
cdf collaboration , phys .
d45 ( 1992 ) 1448 ; m. seymour , z. phys .
c62 ( 1994 ) 127 | we study the production of photons accompanied by jets in large-@xmath0 deep inelastic scattering .
numerical results for the cross section differential with respect to the fraction of momentum @xmath1 carried by a photon inside a jet at large @xmath1 , up to @xmath2 in perturbative qcd , are presented .
the sensitivity to the fragmentation contribution allows one to study the quark - to - photon fragmentation function .
our results can be confronted with future experimental data from hera . |
in current opinion the quark - gluon picture @xcite well describes strong interactions . according to this picture
the mesons are built out of quark ( q ) and antiquark ( @xmath0 ) which are coupled by gluon ( @xmath1 ) exchange .
it is supposed that this model of interaction describes mesons having any signature @xmath2 and lying in any mass region .
this property of description we call _ universality of the quark - gluon picture .
the hypothetic @xmath1 is quark - less , flavor - less electrically neutral particle .
it has property of self - interaction which implies the existence of bound states of two or more gluons .
such an object is called glueball ( g ) and is a singlet state of su(3 ) symmetry .
g has similar properties as the @xmath3 singlet state and can interfere with this singlet as well as with octet isoscalar state , forming in this way new su(3 ) isoscalar states .
_ essential property of g is that it is not observed as a separate particle . to be observable
it must interact with other hadrons . for this to be the case it should have ability to mix with larger unitary multiplet .
being mixed with nonet of @xmath3 states it forms a decuplet @xcite .
then it is subjected to restrictions of unitary symmetry @xmath4 on decuplet components @xcite .
g can be detected as one of the three interfering unphysical components creating the physical isoscalar states of decuplet @xcite .
if it dominates one of these states then it is considered as the `` glueball candidate '' . hence ,
identification of the g is possible on the ground of unitary symmetry .
the unitary symmetry is perceptible due to property of mesons to form multiplets which are collections of particles having different but definite flavors and the same signature @xmath2 .
they form octets , nonets and decuplets .
each multiplet belongs to some representation of unitary symmetry @xmath4 .
the @xmath4 multiplets are analogical to the @xmath5 multiplets describing properties of electromagnetic interaction .
but there is an important difference . in the @xmath5 multiplets differences between masses
are usually neglected .
therefore , this symmetry is considered exact . in the strong interactions
the multiplets gather the particles having _ a priori different masses .
we call them multiplets of broken @xmath4 symmetry .
it is believed that the breaking announces interaction . _ the most apparent effect of breaking @xmath4 symmetry is the difference between isotriplet and isodublet meson masses of the particles belonging to the same unitary multiplet , ( e.g. @xmath6 and @xmath7 mesons ) .
this breaking is described by gell - mann okubo ( gmo ) mass formula for the octet states .
the gmo formula is universal .
other breakings are much weaker @xcite and are masked by gmo one .
they can be exhibited if description of the multiplet complies with gmo requirement .
such a gmo - restricted multiplet is named the multiplet of _ flavor symmetry .
the effect of its breaking is described as _ anomaly .
as gmo breaking is universal the anomalies can be observed in all flavor symmetry multiplets larger than octet .
anomalies of flavor symmetry are very interesting as they comprise information about the interactions which we want to investigate . _
_ anomaly of flavor @xmath4 can be observed as the anomaly of multiplet mass pattern .
also it can be seen due to pattern difference of two ( or more ) identical multiplets differing only in @xmath2 signatures or belonging to various mass regions .
the investigating anomalies provide information on interaction breaking of the @xmath4 symmetry .
we now consider the problem of how to recognize anomalies . in the next two sections we remind how the multiplets are described and define the standard multiplet which provides pattern for anomalies search .
the flavor @xmath4 multiplets are described by the model of vanishing exotic commutators ( vec ) .
the model introduces a system of master equations ( me ) describing the light meson flavor multiplets @xcite @xmath8 @xmath9 where @xmath10 is power index .
symbols of particle @xmath11 , @xmath12 , @xmath7 stand for the mass squared of physical mesons , coefficients @xmath13 express the octet contents of isoscalar states @xmath11 @xmath14 where @xmath13 are real numbers as @xmath15 and @xmath11 describe uncharged states .
hence , @xmath16 and index @xmath17 runs over all isoscalar mesons of the multiplet .
numbering of the isoscalar mesons is chosen so that @xmath17 increases with growing mass : @xmath18 the me are linear with respect to unknown coefficients @xmath19 .
solution defines wave function of the flavor multiplets provided the conditions ( [ d ] ) are satisfied .
the nonet and decuplet are described by the solution of me for the multiplets @xmath20 and @xmath21 , respectively .
the octet masses @xmath12 and @xmath22 make the basic input of me ( [ a ] ) .
there exists only one multiplet which is determined by this input . for other multiplets
this input is insufficient .
the vec description guarantees that all flavor multiplets comply with gmo breaking .
the number of me depends on the multiplet .
the minimal number is @xmath23 , the maximal one is greater by the number of independent mass formulae ( mf ) .
this reflects the fact that , apart from the coefficients @xmath19 , the me determine also the mfs of the multiplet @xcite .
me describe the nonets for @xmath20 .
the vec model reveals some of their properties which were not discussed before .
vec distinguishes three kinds of nonets @xcite : * gell - mann - okubo ( gmo nonet ) having no mf , + * schwinger ( s nonet ) having one mf , + * ideal ( i nonet ) having two mf . the i nonet is described by the solution of the following system of me : + @xmath24 solving the first two equations we calculate @xmath25 , @xmath26 as the functions of masses . next , substituting the @xmath27 s into third and fourth equations we obtain two mfs which determine the masses of mesons @xmath28 , @xmath29 .
the solution of me which describes the i nonet can be given as @xmath30 where n and s are isoscalar states of the ideal nonet .
this solution is an effect of gmo restriction individually acting on @xmath4 nonet states . as the gmo breaking
mechanism is universal the solution ( [ g ] ) is also universal .
the i nonet is the only meson multiplet having this property ; other nonets as well as decuplets are subjected to more complicated breaking .
simplicity of this solution and its explicit connection with gmo - breaking mechanism makes i nonet a pattern enabling exhibition of the nonet anomalies by comparison .
the possibility of confronting properties of different multiplets arise from the fact that they are described by the same vec model .
s nonet is the most frequently appearing multiplet of the light mesons .
it is observed for several @xmath2 mesons in various mass regions @xcite .
the s nonet has one mf . to get the solution for @xmath27 @xmath32 of me we need an extra condition enabling us to evaluate @xmath28 , @xmath29 .
we can use experimental value of one of the masses and calculate the second one with the help of mf .
but the pair of isoscalar meson masses determined in this way differs from the ideal masses ( [ g ] ) resulting from single gmo - breaking of @xmath4 .
the change of s - nonet masses @xmath28 , @xmath29 toward the i - nonet ones discloses the existence of an extra breaking of the already gmo - broken @xmath4 nonet .
that , in turn , suggests the existence of interaction between the i nonet and some unknown particle which causes their mixing and creates a decuplet .
vec predicts two kinds of decuplets : @xmath33 which has no mf and @xmath31 complying with one mf @xcite .
the situation sketched above is described by @xmath31 which is defined by the system of me : @xmath34 the solution determines @xmath19 @xmath35 which define decuplet wave function if the conditions ( [ d ] ) are satisfied .
then the solution defines also the mf of the decuplet @xmath36 the nature of an extra particle whose mixing with the i nonet creates the decuplet is undetermined .
vec only requires that an extra particle extending i nonet to @xmath31 decuplet should be an su(3 ) singlet ; it accepts all _ extending singlets which are usually mentioned , like @xmath37 , hybrid or multi - quark .
then it would be possible that each @xmath31 arises due to mixing with some specific extending singlet and can have individual properties .
_ our purpose is the opposite .
we look for such extending singlet which creates decuplet as universal structure and at the same time describes most of the flavor decuplets . among extending singlets mentioned above only the @xmath37
is supposed to have the property of universality .
therefore , we use this name everywhere .
the masses of the @xmath31 decuplet satisfy the mass ordering rule ( mor ) @xcite @xmath38 mor divides accessible region of the @xmath11 meson appearance into three separate subregions which are delimited by @xmath12 and @xmath22 .
the states in each subregion are uniformly dominated by @xmath39 , @xmath37 , @xmath40 : @xmath41 therefore , we may introduce more transparent notation @xmath42 where n , g , s are basic states of the decuplet isoscalars . + changing indexes one can express mor ( [ ja ] ) in more transparent form @xmath43 where @xmath44 , @xmath45 , @xmath46 are dominated by @xmath39 , @xmath37 , @xmath40 respectively .
this notation is convenient for comparing the properties of the @xmath40 nonet states with the @xmath31 ones .
let us confront s nonet and @xmath31 decuplet descriptions .
+ @xmath40 nonet may be considered as deficient @xmath31 decuplet built on the same basis of octet states @xmath12 and @xmath22 .
mor restrictions on the @xmath40 nonet are known for a long time @xcite .
applying decuplet notation of @xmath39 , @xmath37 , @xmath40-dominated states to nonet isoscalar states @xmath11 we present the nonet mor restrictions as @xmath47 these relations tell us that @xmath37 is included in all @xmath40 nonets .
this suggests its essential role in creating nonets and decuplets .
probably , within these multiplets the suitable @xmath37 is always `` ready to use '' since it is built of the gluons mediating the interaction between quarks included in the multiplet .
this way @xmath37 becomes one of the constituents of the isoscalar mesons .
notice that @xmath48 mesons forming octet include neither scalar @xmath3 state nor @xmath37 one .
observe that the nonet description does not explain the origin of s nonet anomalies .
moreover , the nonets are distinguished by results of biased experiments ignoring possibility of decuplet appearing .
perhaps , results of these measurements should be reanalyzed .
the @xmath31 description looks more reliable .
in such a case the nature of @xmath37 could be probably interpreted in the spirit of van der waals - like interaction .
let us continue the comments .
+ the comparison of the s nonet mor ( [ m]a , [ m]b ) with @xmath31 mor ( [ l ] ) shows that n , g , s - dominated nonet isoscalar states @xmath11 are in both multiplets similarly restricted and occupy the same positions in relation to @xmath12 and @xmath22 bounds .
looking over particle data @xcite we find that the mesons @xmath49 should be g - dominated .
the decay modes of @xmath50 and @xmath51 do not contradict these assignments ; the g dominated structure of @xmath52 meson established earlier @xcite is now confirmed ; the @xmath53 should have g dominated structure if it exists @xcite .
if it is really the case the nature of these mesons have been for a long time notoriously misunderstood .
revealing this misunderstanding must have far - reaching implications to meson spectroscopy .
some of the implications can be seen immediately .
as all s nonets include g - dominated @xmath45 state , it follows that in each of the nonets n - dominated @xmath44 or s - dominated @xmath46 state is missing .
these two omissions reflect two different anomalies of the s nonets .
these nonets describe sets of different particles and implications of anomalies are different .
the axial - vector mesons @xmath54 and @xmath55 provide good illustration of such situation)a and ( [ m])b imply different domains of nonet mixing angle @xmath56 : for ( [ m])a @xmath57 while for ( [ m])b @xmath58 where @xmath59 is ideal mixing angle @xcite . ] .
one of the implications concerns determination of the mass of axial - vector @xmath54 meson @xmath60 .
as s nonet mor does not include the @xmath44 state ( [ m]a ) the @xmath39-dominated state is not represented in this s nonet .
@xmath31 description requires in this region two particles : one isoscalar and one isovector .
this suggests that observed very wide bump of this meson @xcite @xmath61 should be attributed to two states : @xmath62 observe that missing isoscalar state @xmath44 is n - dominated and its structure is similar to structures of the @xmath63 mesons .
this leads to the problem of their separation . among the @xmath55 mesons only the isoscalar @xmath64 and isovector @xmath65 ones
have known masses .
these masses suggest that s nonet of these mesons obeys mor ( [ m]b ) .
according to the decuplet mor ( [ l ] ) the other mesons should satisfy the relations @xmath66 missing one of these particles as well as poor knowledge of remaining particles masses do not allow for further conclusions .
although signals from g dominated and s dominated mesons should be located on opposite sides of @xmath22 it may be difficult to distinguish them .
the attribution of @xmath45 properties ( [ n ] ) to @xmath51 is symbolic .
anomalies of the s nonets provide the evidence for the existence of further ( beyond gmo ) mechanisms of su(3 ) breaking .
the anomalies are caused by interactions which are unknown .
to recognize their nature is just a purpose of the s nonet investigation .
the anomalies provide much weaker breaking than the gmo one @xcite .
this requires much more accurate data to make them observable .
the present data ( partly old and skimpy ) enable completing few nonets but generally are insufficient for completing decuplets .
therefore , in the interest of present and future development of meson spectroscopy it is necessary to increase accuracy of the data and extend the measurements to other @xmath2 mesons .
author thanks professor anna urbaniak - kucharczyk - dean of the physical department of the university of lodz ; professor pawe malanka and professor jakub rembieliski - chiefs of theoretical departments for their support ; special thanks are expressed to professor piotr kosiski for many interesting discussions , valuable comments and reading the manuscript ; dr bartosz zieliski s help in computer manipulations is grateful appreciated . 2 h. fritzsch , m. gell - mann , proc .
the xvi int.conf . on high energy physics , ( * 12 * chicago - batavia ill . ) 1972 vol .
2 , p 135 ; h. fritzsch , p. minkowski , nuovo cim . a * 30*,393 ( 1975 ) ; a. de rujula , h. georgi , s.l .
glashow phys .
d*12 * 147 ( 1975 ) ; m. majewski , w. tybor , acta phys .
b*15 * , 777 ( 1984 ) ; erratum , acta phys . pol .
b*15 * , no 12 page 3 of the cover m. majewski , v. a. meshcheryakov , j. phys .
* 38 * ( 2011 ) 035008 ; m. majewski j. phys .
* 42 * ( 2015 ) 115002 ; m. majewski , w. tybor , acta phys .
b*15 * 267 ( 1984 ) m. majewski , eur .
j. c*30 * , 223 ( 2003 ) ; hep - ph/0206285 particle data group ( revew of particle properties ( 2014 ) ) chinese physics c * 38 * n9 ( 2014 ) 090001 | glueball can be detected as the essential component of light meson decuplet .
the comparison of the decuplet and nonet descriptions of mesons leads to unexpected reinterpretation of the nonet isoscalar states . |
[ [ sketch - of - the - proof - of - the - generality . ] ] sketch of the proof of the generality. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we found the two commuting hamiltonians @xmath39 and @xmath41 in the @xmath42-qubit model depicted in fig .
[ fig : model ] ( example a ) , whose projected counterparts @xmath73 and @xmath44 with the projection ( [ eqn : p1 ] ) of the structure @xmath74 generate @xmath75 .
this single example makes us sure that it is the case for almost all systems . to see this ,
let us formalize in the following way .
take @xmath76 of example a again .
we extract the relevant sector specified by @xmath7 from each element of @xmath29 and call it @xmath77 ( @xmath78 ) , which is a @xmath79 matrix with dimension @xmath80 and is a function @xmath81 of @xmath39 and @xmath41 . together with the @xmath79 identity matrix @xmath82 , the matrices @xmath83 form @xmath84 .
this fact can be mathematically expressed as follows .
we `` vectorize '' each matrix @xmath77 to a @xmath85-dimensional column vector @xmath86 by lining up the columns of the matrix @xmath77 from top to bottom , and gather the column vectors @xmath86 side by side to make up a @xmath87 matrix @xmath88 .
then , the fact that the matrices @xmath83 span @xmath84 is expressed as @xmath89 . note that this determinant is also a function @xmath90 of @xmath39 and @xmath41 . now take a generic couple of commuting hamiltonians @xmath91 and @xmath92 of @xmath42 qubits , i.e. , we randomly choose their eigenvalues @xmath93 , @xmath94 and a common unitary matrix @xmath95 which diagonalizes @xmath91 and @xmath92 simultaneously . inserting this couple of hamiltonians , the determinant @xmath96 is , by construction , a polynomial in the parameters @xmath97 ( @xmath98 ) .
we already know that this polynomial is non - vanishing for the parameter set @xmath99 corresponding to the above specific choice of the hamiltonians @xmath39 and @xmath41 .
therefore , the determinant @xmath100 is a non - zero polynomial in the parameters @xmath97 , implying that its roots are of measure zero in the parameter space . in other words , for almost all parameters @xmath97 , the determinant @xmath100 is non - vanishing , and in turn , almost all couples of commuting hamiltonians become universal , generating @xmath101 , by the projection @xmath7 on the first qubit .
this argument can be generalized to any rank @xmath102 projection , and also to any qubit amplitude damping channel in the strong - damping limit .
the continuous projection @xmath43 required for the qubit - chain model depicted in fig . [ fig : model ] can be induced by an amplitude damping channel acting on qubit 1 .
in fact , consider the master equation @xmath103 with a single lindblad operator @xmath104 which describes the decay of qubit 1 from @xmath105 to @xmath35 , where @xmath35 is associated with the projection @xmath43 in ( [ eqn : p1 ] ) and @xmath105 is the state orthogonal to @xmath35 . solving the system dynamics under the master equation yields @xmath106 + e^{-\gamma t/2}[p_1\rho(0)q_1+q_1 \rho(0 ) p_1]$ ] , where @xmath107 , and @xmath108 represents the partial trace over qubit 1 .
thus , in the limit @xmath109 , we have @xmath110 , and qubit 1 is projected into the state @xmath35 with probability 1 . if this process takes place on a time scale @xmath111 much shorter than any other time scales involved in the dynamics or the controls , then it is effective in inducing a quantum zeno effect on qubit 1 , and it is essentially equivalent to repeating projective measurements .
d. deutsch , `` physics , philosophy and quantum technology . '' in _ proceedings of the sixth international conference on quantum communication , measurement and computing _ , edited by j. h. shapiro and o. hirota ( rinton press , princeton , nj , 2003 ) .
p. zoller , th .
beth , d. binosi , r. blatt , h. briegel , d. bruss , t. calarco , j. i. cirac , d. deutsch , j. eisert , a. ekert , c. fabre , n. gisin , p. grangiere , m. grassl , s. haroche , a. imamoglu , a. karlson , j. kempe , l. kouwenhoven , s. krll , g. leuchs , m. lewenstein , d. loss , n. ltkenhaus , s. massar , j. e. mooij , m. b. plenio , e. polzik , s. popescu , g. rempe , a. sergienko , d. suter , j. twamley , g. wendin , r. werner , a. winter , j. wrachtrup , and a. zeilinger , `` quantum information processing and communication . ''
j. d _ * 36 * , 203228 ( 2005 ) .
k. stannigel , p. hauke , d. marcos , m. hafezi , s. diehl , m. dalmonte , and p. zoller , `` constrained dynamics via the zeno effect in quantum simulation : implementing non - abelian lattice gauge theories with cold atoms . '' _
arxiv:1308.0528 [ quant - ph ] _ | we show that mere observation of a quantum system can turn its dynamics from a very simple one into a universal quantum computation .
this effect , which occurs if the system is regularly observed at short time intervals , can be rephrased as a modern version of plato s cave allegory . more precisely ,
while in the original version of the myth , the reality perceived within the cave is described by the projected shadows of some more fundamental dynamics which is intrinsically more complex , we found that in the quantum world the situation changes drastically as the ` projected ' reality perceived through sequences of measurements can be more complex than the one that originated it .
after discussing examples we go on to show that this effect is generally to be expected : almost any quantum dynamics will become universal once ` observed ' as outlined above .
conversely , we show that any complex quantum dynamics can be ` purified ' into a simpler one in larger dimensions . in the last 30 years
the possibility of using quantum effects to develop an alternative approach to engineering has emerged as a realistic way to improve the efficiency of computation , communication and metrology @xcite . at the very core of this revolutionary idea ,
the possibility of designing arbitrary dynamics of quantum systems without spoiling the rather fragile correlations characterizing them is crucial .
what experimentalists typically do is to apply sequences of control pulses ( e.g. , by sequentially switching on and off different electromagnetic fields ) to steer quantum systems . in the quantum world , however , there is another option associated with the fact that the measurement process itself can induce a transformation on a quantum system . in this context
an intriguing possibility is offered by the quantum zeno effect @xcite . it forces the system to evolve in a given subspace of the total hilbert space by performing frequent projective measurements ( zeno dynamics ) @xcite , without the need of monitoring their outcomes ( _ non - adaptive _ feedback strategy ) .
several attempts have already been discussed to exploit such effects for quantum computation , see e.g. , @xcite . in this work
we show that the constraint imposed via a zeno projection can in fact _
enrich _ the dynamics induced by a series of control pulses , allowing the system of interest to explore an algebra that is _ exponentially larger _ than the original one .
in particular this effect can be used to turn a small set of quantum gates into a universal set .
furthermore , exploiting the non - adaptive character of the scheme , we show that this zeno enhancement can also be implemented by a non - cooperative party , e.g. , by noisy environment . by the zeno effect
, the dynamics of the system is forced to evolve in a given subspace of the total hilbert space @xcite .
one might therefore think that the constrained dynamics is less `` rich '' than the original one .
this naive expectation will turn out to be incorrect .
these surprising aspects of constraints bear interesting similarities to einstein s precepts , according to which one can give a geometric description of complicated motion . the key geometrical idea
is to embed the motion of the system of interest in a larger space , obtaining a forceless dynamics taking place along straight lines .
the real dynamics , with interactions and potentials , is then obtained by projecting the system back onto the original space .
clearly , the constrained dynamics is more _ complex _ than the higher - dimensional linear one . in classical mechanics these reduction procedures , linking a given dynamical system with the one constrained on a lower - dimensional manifold , have been extensively studied as an effective method for integrating the dynamics @xcite . in particular
, different classes of completely integrable systems arise as reductions of free ones with higher degrees of freedom @xcite .
notable examples include the three - dimensional kepler problem , the calogero - moser model , toda systems , kdv and other integrable systems .
the moral is that in classical mechanics , by constraining the dynamics , one often obtains an increase in complexity .
here we find a quantum version of this intriguing effect , which exploits the inherent non - commutative nature of quantum mechanics .
the main idea is that even if two hamiltonians @xmath0 and @xmath1 are commutative , their projected counterparts can be non - commutative @xmath2=0 \quad \not\rightarrow\quad [ php , ph'p]=0,\ ] ] where @xmath3 is a projection . due to this fact
we show that when passing from a set of control hamiltonians @xmath4 to their projected versions @xmath5 one can induce an enhancement in the complexity of the system dynamics which can be exponential , to the extent that it can be used to transform a small number of quantum gates which are not universal into a universal set capable of performing arbitrary quantum - computational tasks .
we find that this effect is completely general and happens in almost all systems .
conversely , we prove that any complex dynamics can be viewed as a simple dynamics in a larger dimension , with the original dynamics realized as a projected dynamics .
what is interesting is that , in contrast to the classical case , the constraint which transforms a hamiltonian @xmath0 into @xmath6 can be imposed not by force but by a simple projective measurement whose outcomes need not be recorded ( the process being effectively equivalent to the one associated with an external noise that is monitoring the system ) .
the underlying mechanism can be rephrased as a modern version of plato s cave allegory @xcite . in the original version of the myth ,
the reality perceived within the cave is described by the projected shadows of some more fundamental dynamics which is intrinsically more complex . in the quantum world , however , the situation changes drastically and the _ projected _ reality perceived within plato s cave can be more complex than the one that has originated it . .
( b ) we perform projective measurements @xmath7 at regular time intervals during the control to check whether or not the state of the system belongs to a given subspace @xmath8 of the global hilbert space .
( c ) in the limit of infinitely frequent measurements ( zeno limit ) , the system is confined in the subspace @xmath8 , where it evolves unitarily with the zeno hamiltonians @xmath9 ( zeno dynamics ) .
the zeno dynamics can explore the subspace @xmath8 more thoroughly than the purely unitary control without measurement . ] [ [ unitary-control-vs.zeno-dynamics . ] ] unitary control vs. zeno dynamics. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in controlled quantum dynamics , two hamiltonians can commute , but their projected versions need not .
this contains , in embryo , the simple idea discussed in the introductory paragraph : interaction can arise from constraints ( in this case projections ) . to describe this mechanism it is worth reminding a few facts about the quantum control theory and the quantum zeno effect . in a typical quantum control scenario
it is assumed that the system of interest ( say the quantum register of a quantum computer , or the spins in an nmr experiment ) can be externally driven by means of sequences of unitary pulses @xmath10 , activated by turning on and off a set of given hamiltonians @xmath11 [ fig .
[ fig : zenodynamics](a ) ] @xcite .
if no limitations are imposed on the temporal durations @xmath12 of the pulses , it is known @xcite that by properly arranging sequences composed of @xmath13 one can in fact force the system to evolve under the action of arbitrary transformations of the form @xmath14 with the anti - hermitian operators @xmath15 being elements of the real lie algebra @xmath16 formed by the linear combinations of @xmath17 and their iterated commutators , @xmath18 $ ] , @xmath19\bm{]}$ ] , etc .
full controllability is hence achieved if the dimension of @xmath20 is large enough to permit the implementation of all possible unitary transformations on the system , i.e. @xmath21 , with @xmath22 being the dimension of the system .
suppose now that between the applications of consecutive pulses @xmath23 we are allowed to perform von neumann s projective measurements [ fig .
[ fig : zenodynamics](b ) ] , aimed at checking whether or not the state of the system belongs to a given subspace @xmath8 of the global hilbert space .
specifically , we will assume that the system is originally initialized in @xmath8 while the various @xmath23 are infinitesimal transformations .
under this condition , the zeno effect can be invoked , in the limit of infinitely frequent measurements , to ensure that with high probability the system will be always found in @xmath8 after each measurement , following a trajectory described by the effective hamiltonians @xmath24 , with @xmath7 the projection onto @xmath8 [ fig .
[ fig : zenodynamics](c ) ] @xcite .
in other words , alternating the control pulses under the frequent applications of the projection @xmath7 the sequence @xmath25 can be effectively transformed into a rotation which on @xmath8 is defined by the unitary operator @xmath26 where @xmath27 .
accordingly the real lie algebra @xmath28 now replaces @xmath20 in defining the space of unitary transformations which can be forced upon the system .
the fundamental result of this paper is to observe that by properly choosing the system setting , the dimension of @xmath29 can be made larger than @xmath30 , to the extent that the former can be used to fully control the system on @xmath8 , in spite of the fact that the latter is not capable of doing the same .
to better elucidate the idea we find it useful to introduce a simple example , where the system is identified with a two - qubit system with control hamiltonians @xmath31 ( we hereafter use @xmath32 to denote pauli operators , and write tensor products as strings , with systems being specified with subscripts and omitting the identity operators ) . notice that their commutator vanishes @xmath33=0 $ ] , and hence the _ naked _ algebra @xmath20 of the two - qubit system has dimension only 2 .
consider now the zeno algebra induced by the projection @xmath34 which freezes the first qubit in the state @xmath35 in the zeno limit .
then , the effective zeno hamiltonians @xmath36 exhibit a non - trivial commutator @xmath37=2ip_{1}y_{2}/3 $ ] , which makes the dimension of @xmath29 equal to 3 ( the situation is schematically illustrated in fig . [ fig : schematic - of - the ] ) .
this in particular implies that @xmath29 can now be used to fully control the system in the subspace @xmath38 ( which is isomorphic to the hilbert space of qubit 2 ) , a task that could not be fulfilled with the original @xmath30 .
[ [ zeno - yields - full - control . ] ] zeno yields full control. + + + + + + + + + + + + + + + + + + + + + + + + + + the example presented in the previous paragraph clarifies that the constrained dynamics can be more complex than the original unconstrained one . the natural question arises : how big can such a difference become ? to what extent can the presence of a measurement process increase the complexity of dynamics in quantum mechanics ? in the following
we provide a couple of examples in which the enhancement in complexity is exponential . while the unprojected dynamics are only two or three dimensional , the projected ones are _
univeral for quantum computation_. this shows that the simple ingredient of projective measurement can strongly influence the complexity of dynamics .
-qubit model described in example a of the text .
straight edges represent the heisenberg interactions , while the triple edge represents the three - body interaction among qubits 13 .
the red part in the upper figure corresponds to @xmath39 acting on qubits 1 and 2 , while the remainder including a local term @xmath40 on qubit 3 corresponds to @xmath41 acting on all the @xmath42 qubits .
the zeno projection @xmath43 on qubit 1 transforms the upper hamiltonians to the lower model , where the state of qubit 1 is frozen , while we are left with a heisenberg chain with the local term @xmath40 and a control @xmath44 on qubit 2 .
the lie algebra of the upper system is only two dimensional , while the lower allows us to perform full control over the system apart from the frozen qubit 1 . ]
example a : consider @xmath42 qubits ( fig .
[ fig : model ] , upper ) , the first two of which are manipulated via the control hamiltonians @xmath45 , and complement it with @xmath41 consisting of the nearest - neighbor heisenberg interactions @xmath46 involving all the qubits but the first two , together with a coupling term acting on the first three qubits and a local term on the third , i.e. , @xmath47 due to the anticommutation of the pauli operators , one can easily verify that the two hamiltonians @xmath39 and @xmath41 commute with each other @xmath33=0 $ ] , defining hence a lie algebra @xmath48 which is barely two dimensional .
now let us consider their constrained versions using the same projection @xmath49 as in ( [ eqn : p1 ] ) . with this choice
we have @xmath50 , and the zeno hamiltonian associated to @xmath51 is given by @xmath52 where now @xmath53 is the nearest - neighbor heisenberg hamiltonian acting on qubits @xmath54 . while qubit 1 is kept frozen in the state @xmath55 by the repetitive projections @xmath43 , the remaining @xmath56 qubits now form a heisenberg chain with a local term on qubit 3 ( fig . [ fig : model ] , lower ) .
elementary but cumbersome calculation shows that with these zeno hamiltonians qubit 2 is fully controllable , which by ref .
@xcite implies that the whole system apart from the frozen qubit 1 is fully controllable .
consequently , we have @xmath57 , so that the zeno algebra is of exponential size , as claimed .
example b : an alternative example which do not involve three - body interactions is available .
consider , for instance , three hamiltonians @xmath58 , @xmath59 , and @xmath60 , and take the zeno projection to be @xmath61 with @xmath43 and @xmath62 projecting qubits 1 and 3 respectively into the states @xmath35 and @xmath63 defined as in ( [ eqn : p1 ] ) .
these hamiltonians commute with each other , and their lie algebra @xmath64 is only three dimensional .
analogously to the previous example , by exploiting the results of ref .
@xcite one can easily show that the dimension of @xmath65 is again exponential , allowing the full control of all the qubits but the first and the third .
[ [ generality - and - hamiltonian - purification . ] ] generality and hamiltonian purification. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + what we have observed above is not a contrived phenomenon , but is actually a quite general one .
considering the couple of hamiltonians @xmath39 and @xmath41 with the projection @xmath43 of the above example a , we are sure that there _ exist _ a pair of commutative hamiltonians and a projection such that the projected dynamics is essentially @xmath66 . a standard argument in control
theory is that if a system is fully controllable for a specific choice of parameters , then it is also fully controllable for almost all parameters @xcite . in our case
it implies that _ almost all _ commuting hamiltonians will become universal through the zeno projection on a single qubit ( see appendix for more details ) .
furthermore , we can show the converse : any non - commutative dynamics can be thought of as the projected version of commutative dynamics in a larger space .
this general phenomenon is in accord with the philosophy of geometrization discussed in the introduction . in analogy with the purification of states in quantum information theory @xcite
, we call it _ hamiltonian purification_. while we give a detailed mathematical analysis elsewhere ,
let us present the simplest case .
consider two arbitrary @xmath22-dimensional hamiltonians @xmath67 and @xmath68 .
we extend the hilbert space by a single qubit and define their `` purifications '' by @xmath69 these extended hamiltonians @xmath39 and @xmath41 are easily seen to commute with each other , @xmath33=0 $ ] , and the projection by @xmath70 yields @xmath71 and @xmath72 , which act as @xmath67 and @xmath68 in the original space before the extension .
we can furthermore apply this procedure iteratively to larger sets of hamiltonians , which means that any complex dynamics can be thought of as a simple one taking place on a larger space , with the complexity arising only from projections .
[ [ local - noise - yields - full - control . ] ] local noise yields full control. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in a classical setting the measurement process is typically perceived as a passive resource that enforces control only when properly inserted in a feedback loop . as explicitly shown by our analysis , and more generally by the results of refs .
@xcite , this is no longer the case in quantum mechanics : measurements can indeed be used to directly drive a quantum system even in the absence of a feedback mechanism .
interestingly enough , for the control scheme we are analyzing here , measurement is not the only way to implement the required projection @xmath7 .
the same effect is attainable by fast unitary kicks and by strong continuous coupling @xcite .
furthermore , owing to the non - adaptive character of the procedure ( we never need to use the measurement outcomes to implement the control ) , it is also achievable by tailoring a strong dissipative process @xcite .
the latter option is of particular interest for us since , along the line of refs .
@xcite , it points out the possibility of taking advantages of the interaction of the system of interest with an external environment , which are typically considered detrimental for quantum processing .
specifically , for the qubit chain analyzed above ( example a ) , one can show that the action of a simple amplitude damping channel @xcite can raise the dynamical complexity to the level of universal quantum computation .
in fact , the decay process bringing qubit 1 to the state @xmath35 can act as a projection @xmath43 ( see appendix ) , and in the strong - damping limit it is effective in inducing a quantum zeno effect on qubit 1 , yielding the full lie algebra @xmath29 in the rest of the qubit chain .
moreover , due to the same reasoning as the one outlined above , almost all qubit amplitude damping channels induce exponential complexity .
[ [ conclusions . ] ] conclusions. + + + + + + + + + + + + + the schemes presented in this work are not meant to be a practical suggestion to implement quantum computers , because the implementation of a control scheme using heisenberg chains would probably be inefficient ( note however @xcite ) .
instead they should be viewed as a proof of the fact that generally adding a simple projection or noise to a dynamical system can profoundly modify the global picture and provoke a drastic _ increase _ in complexity .
this bears some similarities to measurement - based quantum computation @xcite , although there are important differences , in that i ) one does not require the system to be initialized in a complex state , ii ) the measurement is constant , and iii ) its outcome is not used adaptively in future computations @xcite .
our results can be presented as a quantum version of the plato s cave myth , where the projection plays a more active role , making the dynamics of the associated _ quantum shadows _ as complex as universal quantum computation ; and , conversely through hamiltonian purification , a non - commutative dynamics simple .
[ [ acknowledgements . ] ] acknowledgements. + + + + + + + + + + + + + + + + + + this work was partially supported by prin 2010llkjbx on `` collective quantum phenomena : from strongly correlated systems to quantum simulators , '' by a grant - in - aid for scientific research , jsps , by the erasmus mundus - beam program , by a grant for excellent graduate school from the ministry of education , culture , sports , science and technology ( mext ) , japan , and by a waseda university grant for special research projects . |
in the last two decades , considerable progress has been made in loop quantum gravity ( lqg ) , which is a background independent approach to quantum gravity @xcite .
the starting point of lqg is the hamiltonian _ connection dynamics _ of gr rather than the adm formalism . by taking the holonomy of @xmath0-connection @xmath1 and flux of densitized triad @xmath2 as basic variables ,
the quantum kinematical framework of lqg has been rigorously constructed , and the hamiltonian constraint operator can also be well defined to represent quantum dynamics . moreover , a few physically significant results have been obtained in the minisuperspace models of loop quantum cosmology ( lqc ) @xcite . the most interesting one is the _ resolution of big bang singularity _ in lqc @xcite . besides the canonical formalism , the so - called spin foam models were proposed as the path integral formalism of lqg @xcite . however , whether the two approaches are equivalent to each other is a longstanding open question . thanks to the development of lqc , we have a much simpler theory to address this question . as symmetry - reduced models
, there are only finite number of degrees of freedom in lqc . following the conventional method in quantum mechanics
, one can find the path integral formalism of lqc starting with the canonical formalism .
this approach has been implemented by a series of works @xcite with the scheme of simplified lqc @xcite . here
one employed the complete basis of eigen - states of the volume operator to formulate a path integral with somehow _ descrete - steps _ , which inherited certain properties of spin foams @xcite .
moreover , the first - order effective action for the path integral was derived by this approach @xcite , which implied the origin of singularity resolution of lqc in the path integral representation . in canonical lqc , the effective hamiltonian constraint with higher - order quantum corrections could even be obtained by the semiclassical analysis using coherent states , which implied a possible effect of quantum gravity on large scale cosmology @xcite .
it is thus interesting to see whether the effective hamiltonian can be confirmed by some path integral representation .
since the higher - order corrections of the hamiltonian come from the quantum fluctuations , a natural attempt to achieve them is to employ coherent state path integral @xcite . in lqc
, the hamiltonian constraint equation is usually presented to klein - gordon like equation by coupling with a massless scalar field , where the corresponding gravitational hamiltonian operator , as some multiplication of several self - adjoint operators , is non - symmetric in the kinematical hilbert space @xcite . while this treatment is essential in order to obtain the physical states satisfying the constraint equation , it also provides elegant physcial models to examine the so - called _ moyal _ @xmath3__-product _ _ in quantum mechanics . at the very beginning ,
moyal proposed the @xmath3__-product _ _ in order to clarify the role of statistical concepts in quantum mechanics system @xcite .
then this idea were generalized to many situations including quantum spacetime itself .
in canonical quantum theories , the @xmath3__-product _ _ can also be understood by coherent state approach @xcite .
thus it is also possible and desirable to derive the @xmath3__-product _ _ in coherent state functional integral approach .
this idea has been accomplished in the wdw quantum cosmology and lqc @xcite . in these models ,
euclidean _ and _ lorentz _ terms in the gravitational hamiltonian constraint are combined together since they are proportional to each other in spatially flat and homogeneous cases . however
, this is impossible in the full theory , where the lorentz term has to be quantized in a form quite different from the euclidean one @xcite .
this kind of quantization procedure which kept the distinction of the two terms was proposed as alternative dynamics for lqc @xcite .
hence we will study the coherent state functional integral in spatially flat isotropic frw cosmology coupled with a massless scalar field @xmath4 in the alternative quantization framework .
we consider the following hilbert - einstein action of gravity coupled with a massless scalar field : [ eqn : action ] s = r + g^ _ , _ , . in the spatially flat model of frw cosmology , we fix a space - like sub - manifold @xmath5 , which is topologically @xmath6 and equipped with cartesian coordinates @xmath7 and a fiducial flat metric @xmath8 .
the physical 3-metric @xmath9 is then determined by a scale factor @xmath10 satisfying @xmath11 .
it is convenient to introduce an elementary cell @xmath12 and restrict all integrations to this cell in hamiltonian analysis .
the volume of @xmath12 with respect to @xmath8 is denoted as @xmath13 . as in the full loop quantum gravity
, we employ the new canonical variables @xmath14 @xcite . due to the homogeneity and isotropy
, we can fix a set of orthonormal cotriad and triad @xmath15 compatible with @xmath16 and adapted to @xmath12 .
then the basic canonical variables take the simple form @xmath17 , @xmath18 and thus are reduced to @xmath19 with the poisson bracket : @xmath20 , where @xmath21 is the barbero - immirzi parameter . following the _ @xmath22-schem _ of
improved dynamics " @xcite , the regulator @xmath22 used in holonomies is given by @xmath23 , where @xmath24 is a minimum nonzero eigenvalue of the area operator @xcite . in order to do the semiclassical analysis ,
it is convenient to introduce new dimensionless conjugate variables @xcite : [ eqn : b and v ] b:= , v:= , with the poisson bracket @xmath25 , where the planck length @xmath26 is given by @xmath27 . from the matter part of action ( [ eqn : action ] )
, we can get @xmath28 and the poisson bracket : @xmath29 .
the kinematical hilbert space of the quantum theory is supposed to be a tensor product of the gravitational and matter parts . in lqc ,
one employed the standard _ schrdinger _ representation for matter to construct hilbert space @xmath30 , while gravity was quantized by the polymer - like representation @xcite .
thus quantum states in the gravitational hilbert space of lqc are functions @xmath31 of @xmath32 with support on a countable number of points and with finite norm @xmath33 @xcite .
hence the inner product is defined by a _ kronecker delta _ @xmath34 .
the basic operators act on a quantum state @xmath35 in the kinematical hilbert space @xmath36 as : [ eqn : v , b action ] ( v,)=v(v , ) , ( v,)=(v+1 , ) . to obtain the physical states
, one has to solve the quantum hamiltonian constraint equation : [ eqn : quantum constrain lqc ] -(v,)=(-+)(v,)=0 , where @xmath37 is a positive second - order difference operator defined by : ( v,)&= & [ eqn : action te ] , + ( v,)&=&- [ eqn : action tl ] .
here we use the _ alternative quantization _ scheme proposed in @xcite in which the _ euclidean _ and _ lorentz _ terms in the gravitational _ hamiltonian constraint _ are treated separately .
together with the simplified treatment in @xcite , we can get the operators in eqs .
( [ eqn : action te ] ) ( [ eqn : action tl ] ) corresponding to the _ euclidean _ and _ lorentz _ terms respectively .
solutions to the constraint equations and their physical inner products can be obtained through the group averaging procedure .
now we concern about _ coherent state functional integrals_. the ( _ generalized _ ) coherent state of the matter part is labeled by a complex variable @xmath38 and defined by [ eqn : coherent state matter ] : = _ -^ d e^-e^p__o(-_o ) , which is the eigenstate of the _ annihilation _ operator @xmath39 , where @xmath40 describes the width of the wave - packet or quantum fluctuation .
it satisfies the key properties of a coherent state , namely , saturation of heisenberg s uncertainty relation , resolution of identity and peakness property . on the other hand , due to the _ polymer - like _ structure ,
the coherent state of lqc is different from that of the matter part . here
one can define @xmath41 to label the _
generalized _ coherent state @xcite : [ eqn : coherent state gravity ] ( _ _ o:=_v e^-e^-ib_o(v - v_o)(v , where @xmath42 is the characteristic _ width _ of the wave packet and @xmath43 because of the semiclassical feature . for practical use
, one defines the projection of this state on some lattice of variable @xmath32 , saying the _ shadow state _
@xcite : [ eqn : shadow ] ^shad:=_k=-^ e^-e^ib_o(k - v_o ) , k , where we chose the regular lattice @xmath44 .
this shadow state also has the analogous properties of a coherent state .
the resolution of identity now reads [ eqn : gravity identity of coherent ] _
-^dv_o_-^ = _ k=-^ , where the identity @xmath45 is in the subspace in which the states have support only on the regular lattice .
the whole coherent state of lqc reads @xmath46 . in the path integral of the conventional non - relativistic quantum mechanics
, one needs to compute the matrix element of the _ evolution operator _
@xmath47 within the time interval @xmath48 . however , the situation of cosmology in gr is very different , since we are considering totally constrained systems and the operator @xmath49 is not a _ true hamiltonian_. instead , we start from the physical inner product , i.e. , the transition amplitude , of coherent states with _ normalization _ : [ eqn : physical inner coherent ] a([_f],[_i ] ) . to calculate the _ transition amplitude _ , we split a fictitious _ time interval _
@xmath50 into @xmath51 pieces @xmath52 .
to deal with the parameter @xmath53 in group averaging procedure , we employ the trick in @xcite to generalize the one single group averaging to multiple ones , i.e. , & & _ _ o_-_o^_od e^i + & = & _ _ n_o,,_1_o _ -_n_o^_n_od_n _ -_2_o^_2_od_2 _ -_1_o^_1_od_1 e^i(_n++_1 ) , .[eqn : group average n ] in order to trace the power for expansion , we re - scale the parameters by @xmath54 and thus rewrite the exponential operator as : @xmath55 . inserting @xmath51 times of coherent states resolution of identity of @xmath56 and eq . ( [ eqn : gravity identity of coherent ] ) , eq . ( [ eqn : physical inner coherent ] )
can be casted into [ eqn : amplitude2 ] a([_f],[_i ] ) = _ _ n_o,,_1_o _ -_n_o^_n_od_n _ -_2_o^_2_od_2 _
-_1_o^_1_od_1 a^_n a^_n , where a^matt_n=_-^d_n-1 d_1 _ -^ _ n=1^n , [ eqn : a matter ] + a^grav_n=_-^dv_n-1 dv_1 _ -^
_ n=1^n , [ eqn : a gravity ] with @xmath57 .
notice that the characteristic widths @xmath40 and @xmath42 at different steps are not necessarily the same .
so we have to denote @xmath58 and @xmath59 in the semiclassical states @xmath60 and @xmath61 respectively at the n - step " . now
the main task is to calculate the matrix elements of the exponential operators on coherent states .
the exponential operator @xmath62 can be expanded as @xmath63 . for the purpose of a concise writing
, we introduce some intermediate - step notations : _ _
the product of the matrix elements in eq .
( [ eqn : a matter ] ) can be calculated as @xcite : _
n=1^n = ( _ n=1^n ) , [ eqn : a - n matter ] where the product of series @xmath64 can be expressed as & & _
n=1^n = ( _ n=1^n ) + & & .[eqn : prod matter ] here we introduced a _ virtual width
_ @xmath65 by hand , satisfying @xmath66 , in order to get the tidy sum in the exponential position . in the limit of @xmath67 , @xmath65 will approach @xmath68 and hence does not effect the quantum dynamics .
for the gravitational part , careful calculations outlined in the appendix yield & & _
n=1^n = ( _ n=1^n ) + [ eqn : prod lqc result ] where @xmath69 , and & & _
n=1^n = ( _ n=1^n ) + & & .[eqn : prod lqc ] now we take the limit @xmath67 and substitute @xmath70 for @xmath71 to get the functional integral formalism of the amplitude : [ eqn : amplitude result wdw ] a([_f][_i])=e^(|z_f|^2-|z_i|^2+|_f|^2-|_i|^2 ) e^i(s_^matt+s_^grav ) , where [ eqn : effective action matter ] s_^matt=_0 ^ 1d(i ( ) + i()+ + ( p_^2 + ) ) , & & s_^grav=_0 ^ 1d(i ( ) + i()+b + & & + 3 ) . [
eqn : effective action lqc ] here the dots " over @xmath4 and @xmath32 stand for the _ time derivative _ with respect to the _ fictitious time _ @xmath72 .
the _ functional measures _ are defined on _ continuous paths _ by taking the limit of @xmath67 : & & : = _
n _ _ n_o,,_1_o _ -_n_o^_n_od_n _ -_2_o^_2_od_2 _ -_1_o^_1_od_1,[eqn : measure alpha ] + & & : = _ n ( _ n=1^n ) _
n=1^n-1 , [ eqn : measure - matter ] + & & : = _ n ( _ n=1^n ) _
n=1^n-1 .[eqn : measure - grav ] ignoring the total derivatives with respect to @xmath72 in eqs .
( [ eqn : effective action matter ] ) and ( [ eqn : effective action lqc ] ) , we can read out the total effective hamiltonian constraint as : _
eff=-- & -&3g^2 + & & + .[eqn : effective constraint lqc ] note that @xmath73 and @xmath74 are the square of fluctuations of @xmath75 and @xmath76 respectively .
they can be seen as quantum corrections to the leading term : @xmath77 of the _ euclidean _ part @xmath78 as well as the leading term : @xmath79 of the _ lorentz _ part .
a careful observation reveals that the real and imaginary parts of the leading terms can be synthesized into a _
moyal _ @xmath3-_product _
@xcite , i.e. , & & v^2 ^ 2(2b)+iv = ve^(_b -_v ) ( ) = : v*()[eqn : lqce*prod ] , + & & v^2 ^ 2(4b)+2iv = v*()[eqn : lqc*prod ] .
therefore the effective hamiltonian constraint takes the form : & & _
eff=-- -3g^2 + & & + .
[ eqn : effective constraint lqc with * ] the _ moyal _ @xmath3-_product _ emerges in the gravitational part of the hamiltonian , since both @xmath80 and @xmath81 are non - symmetric operators which can be regarded as a product of two self - adjoint operators .
thus , in this model the coherent state functional integral also suggests the _ moyal _ @xmath3-_product _ to express the effective hamiltonian for the quantum system with a _ non - symmetric _ hamiltonian operator .
however , since the _ moyal _ @xmath3-_product _ originates from the non - symmetry of the operator , one may doubt why we did not use a _ symmetric _ operator from the very beginning . to understand the motivation of the non - symmetric operator @xmath82 , we recall that the initial hamiltonian constraint operator in lqc is actually self - adjoint in the kinematical hilbert space@xcite . to resolve the constraint equation and find _ physical _ states ,
one feasible method is to rebuild the constraint equation as a _ klein - gordon _ like equation and treat the scalar @xmath4 as an _
internal time_. as a result , the constrained quantum system was recast into a relativistic particle whose dynamics is govern by a _
klein - gordon _ like equation with an _ emergent time _ variable @xcite .
the price to get this klein - gordon like equation is that the new gravitational hamiltonian operator @xmath82 becomes a multiplication of two self - adjoint operators , and hence it is no longer symmetric .
but this does not indicate that one could not employ @xmath82 in the intermediate step to find physical states .
on the other hand , because the _ moyal _ @xmath3-_product _ comes from the expectation value of the _ multiplication _ of two self - adjoint operators on coherent state , this non - symmetric @xmath82 just provides a suitable arena to examine the _ moyal _ @xmath3-_product _ from the path integral perspective .
we can also take another practical way to symmetrize @xmath82 at the beginning .
for example , one can define a symmetric version of @xmath78 by ^:=(+^ ) , and then carry out the same procedure of above coherent state functional integral . in the calculation of matrix element @xmath83 , we could think that the operators @xmath75 and @xmath84 in @xmath78 act on _ bra _ @xmath85 and _ ket _ @xmath86 respectively , while @xmath84 and @xmath75 in @xmath87 act on _ bra _ @xmath85 and _ ket _ @xmath86 respectively .
then it is not difficult to see that the imaginary parts generated by @xmath78 and @xmath87 cancel each other .
hence for the symmetric hamiltonian operator corresponding to @xmath82 , we can finally get the following effective hamiltonian constraint : [ eqn : effective constraint lqc remove * ] : = --+3(v^2 + ) ( ^2(2b)(1-(16 + 12 ^ 2)^2 -(1+^2)(1 - 16 ^ 2)^2(2b))+2 ^ 2 ) , which takes the same form as ( [ eqn : effective constraint lqc with * ] ) but without @xmath3-_product _ while @xmath88 denotes the quantum fluctuation of @xmath89 .
using the effective hamiltonian constraint @xmath90 which contains _ moyal _ @xmath3-_product _ , one may investigate the corresponding dynamics by defining the evolution equation as : [ eqn : evolution with * ] ( v , b):=(f*_eff-_eff*f ) , for any dynamical quantity @xmath91 .
especially , the evolution of basic variables can be obtained as : & = & - + & & + , [ eqn : v evolution with * ] + & = & + & & - [ eqn : b evolution with * ] where @xmath92 and @xmath93
. however , there seems no way to understand eqs .
( [ eqn : v evolution with * ] ) and ( [ eqn : b evolution with * ] ) directly as effective classical equations because of the @xmath3-_product _ therein . instead , we could use the effective hamiltonian constraint ( [ eqn : effective constraint lqc remove * ] ) without @xmath3-_product _ to explore the effective dynamics . using the conventional poisson bracket
, we can derive a modified friedmann equation from the effective hamiltonian ( [ eqn : effective constraint lqc remove * ] ) as : & & h^2()^2 = ^2[eqn : fdm eq ] where @xmath94 is a constant . to annihilate @xmath95 and @xmath96 in eq .
( [ eqn : fdm eq ] ) , we use the constraint equation ( [ eqn : effective constraint lqc remove * ] ) to get [ eqn : sin and rho ] ^2(2b)= , and [ eqn : chi ] -2 ^ 2 , where @xmath97 is the density of matter , @xmath98 , and @xmath99 . however , eq . ( [ eqn : fdm eq ] ) looks problematic since it depends on the volume @xmath32 of the chosen fiducial cell .
this originates from the fact that we have to use the coherent states peaked on the phase points @xmath100 in the path integral . in the final picture
we have to remove the infrared regulator by letting the cell occupy full spatial manifold . in this limit ,
the irrelevant correction terms proportional to @xmath101 could be neglected , while the relevant terms proportional to @xmath102 would be kept , since @xmath103 was understood as the fluctuation of @xmath89 which does not depend on the fiducial cell .
we finally get & & h^2=^2[eqn : fdm eq2 ] where the positive and negative signs correspond to the expanding and contracting universe respectively . here
we use notations : @xmath104 and @xmath105 for a concise writing . note that eq .
( [ eqn : fdm eq2 ] ) implies significant departure from classical gr . for simplicity
, we first consider the case @xmath106 and see whether the _ bounce _ or _ re - collapse _ determined by @xmath107 could occur .
then it is obvious that , for a contracting universe , the so - called quantum bounce of lqc will occur when [ eqn : solution chi 1 ] = , which means that @xmath108 increases to @xmath109 if @xmath110 is neglected . on the other hand , for an expanding universe
, a re - collapse would occer when @xmath111 , or equivalently @xmath108 decreases to @xmath112 , which coincide with the result in canonical theory @xcite , where @xmath113 was assumed as higher order term and hence neglected . as pointed out in refs.@xcite ,
the inferred re - collapse is almost in all probability as viewed from the parameter space characterizing the quantum fluctuation @xmath114 .
intuitively , as the universe expands unboundedly , the matter density would become so tiny that its effect could be comparable to that of quantum fluctuations of the space - time geometry .
then the hamiltonian constraint may force the universe to contract back .
it should be noted that the effective equation and hence the inferred effect of re - collapse are only valid with the coherent states of gaussian type . whether there is a similar result for other semiclassical states is still an interesting open issue .
for example , one may consider the affine coherent states in the _ affine quantum gravity _
approach developed by klauder @xcite . in the case when @xmath113 could not be neglected , the bounce would also be approached for a contracting universe . because @xmath113 would not be bigger than the order of @xmath115 , @xmath116 could be infinitely close to the result in eq .
( [ eqn : solution chi 1 ] ) and lead to @xmath107 .
however , for an expanding universe , both of the two terms in the bracket of the right hand side of eq .
( [ eqn : fdm eq2 ] ) are non - negative in large scale . as a result ,
the hubble parameter would always keep non - zero unless @xmath117 approaches @xmath118 asymptotically .
therefore , the inferred re - collapse might occur only if @xmath117 approaches @xmath118 asymptotically . if we neglect all the higher - order quantum corrections : @xmath119 , @xmath110 and @xmath113 , eq .
( [ eqn : chi ] ) would be simplified to @xmath120 , and hence a _ first - order _ modified friedmann equation could be obtained from eq .
( [ eqn : fdm eq2 ] ) as : [ eqn : fdm eq3 ] h^2= .
note that this first - order modified friedmann equation is different from that in refs.@xcite .
but it coincides with the modified friedmann equation in ref.@xcite . hence eq .
( [ eqn : fdm eq3 ] ) still contains the particular information of alternative dynamics .
it is easy to see that if the matter density increase to @xmath121 , hubble parameter would be zero and the _ bounce _ could occur for a contracting universe . on the other hand , in the classical regime of large scale
, we have @xmath122 for @xmath123 and hence eq .
( [ eqn : fdm eq3 ] ) reduces to the standard classical friedmann equation : @xmath124 .
since there are quantization ambiguities in constructing the hamiltonian constraint operator in lqc , it is crucial to check whether the key features of lqc , such as the quantum bounce and effective scenario , are robust against the ambiguities . moreover , since lqc serves as a simple arena to test ideas and constructions induced in the full lqg , it is important to implement those treatments from the full theory to lqc as more as possible . unlike the usual treatment in spatially flat and homogeneous models , the lorentz term has to be quantized in a form quite different from the euclidean one in full lqg . for the above purpose ,
this kind of quantization procedure which kept the distinction of the lorentz and euclidean terms was proposed as alternative dynamics for lqc @xcite .
it was shown in the resulted canonical effective theory that the classical big bang is again replaced by a quantum bounce and it is possible for the expanding universe to re - collapse due to the quantum gravity effect by certain assumption .
hence it is desirable to study such kind of predictions from different perspective .
meanwhile , it is also desirable to study the _ moyal _ @xmath3-_product _ by coherent state functional integral approach within lqc models . to carry out the above ideas , the present paper is devoted to study the coherent state functional integral in spatially flat isotropic frw model coupled with a massless scalar field in the alternative dynamics framework of lqc .
the main results can be summarized as follows . by the well - established canonical theory ,
the coherent state functional integral for lqc with alternative dynamics has been formulated by group averaging . for the non - symmetric gravitational hamiltonian constraint operator ,
the _ moyal _ @xmath3-_product _ emerges naturally in the resulted effective hamiltonian with higher - order quantum corrections . for the corresponding symmetrized hamiltonian operator , the effective hamiltonian and modified friedmann equation
are also derived from the coherent state functional integral approach .
it turns out that the quantum bounce resolution of big bang singularity can also be obtained by the path integral representation . on the other hand ,
if higher order corrections are included , there is a possibility for the re - collapse of an expanding universe due to the quantum gravity effect , which coincides with the result obtained in the canonical formalism .
moreover , the first - order modified friedmann equation still contains the particular information of alternative dynamics and hence admits the possible phenomenological distinction between the different proposals of quantum dynamics .
the alternative modified friedmann equation ( [ eqn : fdm eq2 ] ) or ( [ eqn : fdm eq3 ] ) sets up a new arena for studying phenomenological issues of lqc .
we thank abhay ashtekar and peng xu for helpful suggestion and discussion .
this work is supported by nsfc ( no.10975017 ) and the fundamental research funds for the central universities .
we give some details on the calculation of the lorentz part of the matrix element of exponentiated gravitational hamiltonian operator : @xmath125 .
the order of @xmath126 of this matrix element is & = & i _ k + & & i ( l^+_n , n-1-l^0_n , n-1+l^-_n , n-1).[eqn : app - matirx - element ] now we need to deal with the three terms @xmath127 separately .
first , we get l^+_n , n-1&&_k(k^2 + 4k)e^-- e^-ib_n(k - v_n)+ib_n-1(k+8-v_n-1 ) + & = & + & & _ k(k^2 + 4k),[eqn : d+ ] where @xmath128 .
to do the summation in the above equation , we first have to rewrite @xmath129 as a function of @xmath130 : k^2 + 4k&=&(k-^+_n)^2+(2^+_n+4)(k-^+_n ) + ^+_n^2 + 4^+_n + & = & ( k-^+_n)^2 + 2(_n-)(k-^+_n ) + _ n^2-_n + , where @xmath131 . do the summation term by term , we get l^+_n , n-1= e^-e^i8_n ( ( _ n)^2-_n+ + p^+_n , n-1 ) , where @xmath132 denotes a polynomial of @xmath133 , @xmath134 and @xmath135 without the zeroth order term . here
we have expanded the factor @xmath136 in eq .
( [ eqn : d+ ] ) as @xmath137 . except for the leading term @xmath138
, all the other terms can be conflated with @xmath132 . under the continuous limit @xmath139
, this @xmath132 does not contribute to the effective action of gravity .
it is easy to calculate @xmath140 and @xmath141 as follows : & & l^0_n , n-1=2 ( ( _ n)^2 + + p^0_n , n-1 ) , + & & l^-_n , n-1= e^-e^-i8_n ( ( _ n)^2+_n+ + p^-_n , n-1 ) . taking the expansion @xmath142 and neglecting the higher order terms than @xmath143 , we can get the combination & & l^+_n , n-1-l^0_n , n-1+l^-_n , n-1 + & = & -4 , and hence the matrix element @xmath125 is .&&[eqn : app - l - element ] similarly , we can get the euclidean part @xmath144 as : = .&&[eqn : app - e - element ] combining eqs .
( [ eqn : app - l - element ] ) and ( [ eqn : app - e - element ] ) , we can get eq .
( [ eqn : prod lqc result ] ) . | coherent state functional integral for the minisuperspace model of loop quantum cosmology is studied . by the well - established canonical theory ,
the transition amplitude in the path integral representation of loop quantum cosmology with alternative dynamics can be formulated through group averaging .
the effective action and hamiltonian with higher - order quantum corrections are thus obtained .
it turns out that for a non - symmetric hamiltonian constraint operator , the _ moyal ( star)-product _ emerges naturally in the effective hamiltonian . for the corresponding symmetric hamiltonian operator ,
the resulted effective theory implies a possible quantum cosmological effect in large scale limit in the alternative dynamical scenario , which coincides with the result in canonical approach .
moreover , the first - order modified friedmann equation still contains the particular information of alternative dynamics and hence admits the possible phenomenological distinction between the different proposals of quantum dynamics .
+ * keywords * : loop quantum cosmology ; coherent state functional integral ; effective theory |
rare events are physical events occurring with low probability in numerous many - body systems , encompassing a wide range of phenomena , from condensed matter physics to proteins : they often involve thermally activated processes , i.e. passages between different stable configurations of the system that are separated by free energy barriers . evaluating the frequency of these rare events from numerical simulations
can be very time consuming , as the probability of observing one of these passages is very low : for instance , the migration of a vacancy in @xmath1-iron at 500k typically happens every microsecond , while the usual time steps for molecular dynamics is of the order of a few femtoseconds . in the last decades ,
several techniques have been developed in order to accelerate the dynamics , and enhance the probability of observing infrequent events during a short simulation , based on importance sampling .
@xcite among these techniques , transition path sampling @xcite ( tps ) allows to estimate the frequency of rare events by means of path ensemble averages of physical observables : reaction rates for appropriate time scales are indeed evaluated from the ratio of the number of reactive trajectories on the total amount of paths sampled .
however , a sufficient number of reactive paths has to be observed , in order to obtain a reliable statistics .
herein , we propose a transition path sampling algorithm where the fraction of reactive paths sampled is enhanced using an adequate bias that favors diverging trajectories .
it is indeed possible to show @xcite that reactive paths we want to sample share important features with diverging trajectories observed in chaotic systems .
therefore , a suitable parameter that quantifies the divergence of dynamic trajectories can be exploited also to bias a path sampling algorithm aimed to reproduce reactive paths .
the main instrument proposed in the literature to quantify chaotic behavior of dynamical systems is the evaluation of lyapunov exponents , @xcite that are usually employed to estimate the sensitivity of deterministic systems to small changes in initial conditions . for this feature ,
they have been widely studied , @xcite both analytically and numerically , in hamiltonian as well as in nonlinear systems of small dimensions . moreover
, the use of lyapunov exponents to characterize numerically phase transitions in finite size systems has been extensively explored in the past years , and many noticeable results have been obtained in the early 90 s .
@xcite resorting to lyapunov exponents in order to achieve numerical ergodicity and localize saddles and transition paths has been done recently with the lyapunov - weighted dynamics method proposed by tailleur and kurchan : @xcite in this sampling scheme , a set of clones are copied or deleted depending on a probability weight computed from quantities related to lyapunov exponents .
after this work , the paper of geiger and dellago @xcite have shown how to couple the chaoticity features of a dynamical system to a tps technique for sampling deterministic trajectories , using an indicator for diverging trajectories borrowed from studies on planetary systems @xcite , the relative lyapunov indicator ( rli ) . in this paper , we present a chaoticity indicator different from rli , and based on local lyapunov numbers , that are quantities closely related to lyapunov exponents .
this indicator is used to introduce a bias in the path sampling scheme , thus obtaining a lyapunov biased tps ( lytps ) method that will be in the sequel applied to complex many - body systems , like the well known optimization benchmark model lj@xmath0 .
furthermore , we show how reaction rate constants can be recovered from biased tps quantities , resorting to an appropriate statistical analysis to unbias reaction rate values computed in a lytps framework .
this paper is organized as follows . in sec .
[ sec : lyapunov - exponents - in ] , we first recall the basic concepts of lyapunov exponents for dynamical systems . we will then briefly review the use that has been made of them in numerical algorithms to characterize phase transitions , or in importance sampling contexts .
we then expose how to use local lyapunov numbers in the context of a transition path sampling to determine saddle points and reactive paths ( sec .
[ sec : transition - path - sampling ] ) .
reaction constants are computed from the fraction of reactive paths using the bennett - chandler approach @xcite of population correlation functions and the standard tps technique @xcite ; we also explain how unbiased reaction constants are recovered from a lyapunov biased algorithm thanks to the multistate bennett acceptance ratio @xcite ( mbar ) method ( sec .
[ sec : reaction - rate - constants - calculation ] ) .
finally , numerical results concerning the application of our method to solid - solid structural transition in lj@xmath0 and vacancy migration in @xmath1-iron are presented ( sec .
[ numerical results ] ) .
we briefly present here the theory of lyapunov exponents , mainly following ott @xcite ; then we propose a formulation allowing the use of these exponents in importance sampling techniques . in numerical applications ,
dynamics are discrete : the evolution of state vector @xmath2 at time step @xmath3 is described by a _ mapping _
@xmath4 , where @xmath5 is a matrix expressing the system evolution from one time step to the following .
let the system be at time @xmath6 in an initial position @xmath7 , and let @xmath8 be a small perturbation applied to this initial state .
the dynamics of such a perturbed system can then be denoted using a new state vector @xmath9 , whose time evolution will be @xmath10 for a sufficiently small perturbation @xmath11 , it is possible to linearize the map @xmath5 as @xmath12 the time evolution of a small perturbation to the initial state vector then reads @xcite @xmath13 where @xmath14 is the jacobian matrix of the map @xmath5 . inserting particular solutions @xmath15^{n}$ ] in eq . , one finds an eigenvalue equation @xmath16 in the discrete case , the eigenvalues @xmath17 of @xmath18 , solutions of eq . ,
are called _ lyapunov numbers _ rather the lyapunov exponents , and trajectories are _ unstable _ for @xmath19 and stable otherwise .
the largest eigenvalue @xmath20 will be associated to the eigenvector @xmath21 indicating the direction of maximal growth of the perturbation @xmath11 .
we introduce the matricial product @xmath22 between the jacobian matrices of the hamiltonian map at successive time steps , and we express the perturbation at time step _
n _ with respect to the initial perturbation @xmath8 as @xmath23 defining with @xmath24 the euclidean norm of @xmath8 in phase space , the _ lyapunov exponents _
@xmath25 , given the initial condition @xmath7 and the initial perturbation orientation @xmath26 are @xmath27 for _ n_-body hamiltonian systems , having _
3n _ degrees of freedom , @xmath28 is a _ 6n_-dimensional state vector accounting for both the positions and momenta of the @xmath29 particles .
for the _ 6n_-dimensional hamiltonian map there will be _
6n _ lyapunov exponents , usually ordered in literature from the largest to the smallest ( @xmath30 ) .
as stated in ref .
@xcite , oseledec s multiplicative ergodic theorem @xcite guarantees the existence of the limits used in the definition of the lyapunov exponents in eq . under very general circumstances .
the lyapunov exponents are related to the aforementioned lyapunov numbers as @xmath31 $ ] . from eq .
, we can also define _ finite - time lyapunov exponents _ :
@xmath32 for long enough times , the greatest lyapunov number @xmath33 will give the dominant contribution to the perturbation evolution , and the associated eigenvector @xmath34 indicates the direction of maximum growth of the perturbation @xmath11 .
we stress here that finite - time lyapunov exponents are calculated for a given @xmath7 and that , strictly speaking , their values do depend on the initial orientation @xmath35 .
it is however shown that the largest exponent @xmath36 is approximately independent of the choice of @xmath35 in hamiltonian ergodic systems@xcite , while the complete spectra of finite - time exponents can be determined using specific numerical techniques .
@xcite in numerical simulations , only finite - time lyapunov exponents can be estimated , due to limited amount of cpu time .
to evaluate @xmath37 from eq .
we should compute the matricial product of eq . . for systems with many degrees of freedom ,
a calculation of this matrix product is not possible analytically , and is numerically expensive .
the solution most followed in literature consists in directly evaluating the quantity @xmath38 given by the distance @xmath39 between two nearby dynamical trajectories , the first one started from the initial state @xmath7 and the second one from the perturbed configuration @xmath40 after _
n _ time steps . the use of eq . as a mean to evaluate finite - time lyapunov exponents has two main drawbacks : the need of computing two trajectories to evaluate a single lyapunov exponent , thus doubling computational cost , and the fact that values obtained for @xmath41 can be sensitive to initial conditions , because of the dependence of the computed finite - time lyapunov exponents from the choice of the orientation of initial perturbation @xmath35 , as recalled above .
several numerical strategies have been proposed to bypass these two problems .
the tangent space method @xcite assigns to each state @xmath42 of the trajectory started in @xmath43 a vector @xmath44 .
these vectors are computed from the local hessian matrix of the hamiltonian mapping , and their norms indicate the distance between the current trajectory and the perturbed one , i.e. @xmath45 . as these distances evolve exponentially ( see eq . ) , the lengths of the vectors @xmath46 can quickly diverge or vanish : a reorthonormalization of @xmath46 at each time step is therefore required , for instance with a gram - smith algorithm .
this method has been implemented in the literature @xcite , for instance in the context of lyapunov weighted dynamics .
@xcite to make this algorithm independent of the choice of the first vector @xmath47 , one could integrate the equations of motion backward in time from @xmath43 for a duration @xmath48 , and then reintegrate the evolution of @xmath44 forward until @xmath6(see ref .
@xcite ) . in this way
, @xmath47 would be automatically oriented in the direction of maximum growth .
however , the duration @xmath48 should be long enough to ensure the loss of correlation between the orientation of @xmath49 and @xmath47 , thus requiring the computation of long trajectories at sustained computational cost .
@xcite another evaluation of finite - time lyapunov exponents can be obtained with the relative lyapunov indicator ( rli ) , elaborated by sndor _
_ @xcite in the context of planetary trajectories , and further used in a lyapunov weighted path sampling scheme @xcite .
the main idea is to compare finite - time lyapunov exponents @xmath41 for trajectories starting very close , say in @xmath7 and @xmath50 .
the difference between finite - time exponents at time step @xmath3 can be written as @xmath51 and will in general undergo strong fluctuation , @xcite which can be smoothed by an average over @xmath52 trajectory steps : in this way one defines the rli as the quantity @xmath53 this average over the entire trajectory length reduces @xcite the dependence of the computed finite - time lyapunov exponents on the orientation of initial perturbation , but introduces an additional dependence on @xmath54 .
both finite - time lyapunov exponents are calculated evaluating the distance between two trajectories evolving close to each other ( instead of the matricial product of eq . ): in terms of computational cost , four trajectories are computed to obtain a single rli . in the following ,
we propose a faster and orientation - independent way to evaluate the divergence of hamiltonian trajectories , alternative to tangent space method and rli , to be used in the path - sampling scheme described in sec . [
sec : transition - path - sampling ] .
we restrict our focus to systems with deterministic dynamics governed by an hamiltonian @xmath55 .
the time evolution of the state vector @xmath56 ( also indicated as `` hamiltonian flow '' @xcite ) directly follows from hamilton equations , @xmath57 the evolution of a small perturbation of positions and momenta for a standard hamiltonian is then given by @xmath58 where @xmath59 is the @xmath60 inverse mass matrix of elements @xmath61 .
let us discretize the hamiltonian dynamics in eq . with the velocity verlet
algorithm : @xmath62 this algorithm is accurate to second order and numerically stable @xcite
. it will be used in sec .
[ sec : transition - path - sampling ] to generate dynamical trajectories in numerical applications .
the perturbation @xmath63 can be evaluated with respect to @xmath64 using eq . , where the jacobian matrix of the hamiltonian mapping for a velocity verlet discretization reads @xcite @xmath65\right\ } & \mathbf{i}-\frac{dt^2}{2m}\mathbf{h}(\mathbf{x}_{n+1})\end{array}\right]\]]in the upper right and bottom left blocks we introduced the hessian matrix of the potential energy @xmath66 at states @xmath67 and @xmath68 , respectively . the jacobian matrix of eq . obtained with the velocity verlet scheme of eq
contains the hessian matrices at steps @xmath3 and @xmath69 .
therefore , manipulating @xmath14 is numerically expensive .
a simpler expression for the jacobian matrix can be obtained from the less accurate euler discretization algorithm : eq .
becomes a set of _
6n _ coupled equations of motion @xmath70 where @xmath71 , so that the jacobian matrix of the hamiltonian map reads @xmath72.\ ] ] again , the perturbation @xmath73 at each time step can be evaluated by inserting eq . in eq . .
the difference between the jacobian matrix @xmath14 of eq .
and the one of eq .
consists in second order terms .
however , it is numerically less expensive to manipulate the former than the latter , as @xmath14 of eq .
requires to evaluate the hessian only at time step @xmath3 . in the following , we will be interested in computing the eigenvalues of @xmath14 , in order to obtain a bias favoring reactive trajectories : this bias will be removed at the end , so it would be useless to spend cpu time to accurately evaluate the jacobian matrix . therefore , accordingly to ref .
@xcite , we consider the euler scheme ( eq . ) precise enough for our purposes , and we use the first order euler discretization of eq . to compute @xmath14 . using the discretized hamiltonian dynamics given in eq
, we proceed by computing at each time step the maximum local lyapunov number @xcite , given by the largest eigenvalue of @xmath74 ( eq . ) .
the @xmath75 eigenvalues @xmath76 of @xmath74 , computed at time step @xmath3 , can be obtained writing eq . as a secular equation @xmath77 whose solutions are @xmath78 pairs of eigenvalues @xmath76 of @xmath14 , because of the simplectic properties of the hamiltonian mapping matrix @xmath5 .
@xcite these eigenvalues are given by the expression @xcite @xmath79 where the @xmath80 correspond to the eigenvalues of the mass - weighted hessian @xmath81 of components @xmath82 eq .
shows that at each time step @xmath3 the jacobian eigenvalues @xmath83 , i.e. the local lyapunov numbers , depend on the potential energy surface through the hessian eigenvalues @xmath80 : unstable configurations @xmath84 , such as saddle points are characterized by the occurrence of some negative @xmath80 , and correspond to real and positive local lyapunov numbers @xmath83 .
conversely , stable states have all @xmath80 positive or imaginary local lyapunov numbers with unitary real part . in the following path sampling scheme of sec .
[ sec : transition - path - sampling ] we neglect the imaginary part of @xmath83 given by stable states .
this is not an issue because it is sufficient to determine real and positive lyapunov exponents to characterize unstable dynamics which we are interested in , and the imaginary part of the jacobian eigenvalues @xmath83 can be neglected . indeed , the imaginary part of the lyapunov exponents does not affect the stability of the system , but only indicates if @xmath85 is spiraling clockwise or counterclockwise . at each time step @xmath3 ,
the most negative eigenvalue of the hessian matrix @xmath86 gives the eigenvalue of the jacobian matrix @xmath14 with the largest real part .
we write the maximum local lyapunov number as @xmath87 such that eq
. entails @xmath88 for stable configurations , where all @xmath89 are positive , and @xmath90 for unstable configurations having a negative spectra .
eigenvalues @xmath91 are found by diagonalization of the hessian matrix @xmath92
. this can be computationally very expensive for systems with a large number of degrees of freedom .
one efficient solution consists in extracting only the lowest eigenvalue @xmath93 needed to evaluate @xmath94 using the lanczos algorithm @xcite .
this iterative algorithm finds extremal eigenvalues of any matrix with a reduced computational cost , diagonalizing only a submatrix of the initial one ( see for example appendix a of ref .
@xcite for details ) . as pointed out in ref .
@xcite , a @xmath95 lanczos submatrix is sufficient to detect negative eigenvalues .
moreover , it is possible to decrease the submatrix size to as little as @xmath96 by verifying at each iteration that the lanczos solution is stable ; if not , repeat the calculation until a the solution is converged .
hence , the most negative eigenvalue , corresponding to the most unstable direction of the potential energy surface at a give system position in the phase - space at a given instant can be extracted in a very few iterations .
this is the numerical method we will apply in the following to evaluate @xmath94 .
a dynamical trajectory is defined as an ordered sequence of states in phase space separated by a small time increment @xmath97 , and denoted as @xmath98 , i.e. a path of total length @xmath48 composed by @xmath99 state vectors .
we introduce a lyapunov indicator for path @xmath100 @xmath101 given by the average of the maximum lyapunov number of eq over the whole trajectory . the verification of the difference between finite - time lyapunov exponents estimated by rli or tangent space method and @xmath102 from eq .
is beyond the scope of this article .
we stress instead that the indicator proposed in eq .
, being based on the hessian spectra @xmath89 , is strictly related to the topological properties of the potential energy surface , thus encodes local information on the stable or unstable configurations sampled in phase space by a given trajectory .
henceforth , we consider this lyapunov indicator suitable for importance sampling techniques .
the largest local lyapunov number of eq . were evaluated by hinde _
et al . _
@xcite to study the dependence of the kolmogorov entropy on the potential energy surface of small lennard - jones clusters . in that work ,
the lyapunov exponents derived from the eigenvalues of the jacobian matrix of the hamiltonian mapping are summed over trajectories of different lengths , thus obtaining - using pesin s theorem @xcite - an estimation of kolmogorov entropy . using eq . to evaluate finite - time lyapunov exponents
was proven to be a quite successfull approach , as it yields enough information to quantify the degree of instability of phase space trajectories .
this supports the idea that evaluating the global lyapunov exponent from local lyapunov numbers allow to correctly characterize the chaotic properties of the system .
the idea of sampling the phase space of a many - body system through paths generated by molecular dynamics , using the metropolis algorithm , has been developed by dellago and coworkers .
@xcite this approach is called transition path sampling ( tps ) . in this present work ,
we introduce a bias in the tps algorithm in order to favor the sampling of reactive trajectories .
the bias is proportional to the lyapunov indicator @xmath103 obtained in eq . .
each path @xmath104 constrained to start in a reactant basin a is equipped with a probability density @xmath105 where @xmath106 is the partition function on the biased trajectory ensemble , and function @xmath107 is an additional term linking the initial state @xmath7 of the path to state @xmath108 .
different choices for @xmath109 are possible , for instance @xmath110 , where @xmath111 is an indicator function on a , such that @xmath112 or @xmath113 that accounts for having a tunable spring of stiffness @xmath114 linking the origin of path @xmath104 to state a. in this last case , the stiffness parameter @xmath114 can be tuned to counterbalance the strenght of the bias .
we denoted in eq . for simplicity @xmath115\ ] ] as the dynamical path probability arising from the deterministic propagation of the trajectory .
@xmath116 is the unnormalized canonical distribution at inverse temperature @xmath117 from which the initial configuration is selected , and @xmath118 is the _ temporal propagator _
@xmath119 associated to the deterministic dynamics @xcite .
in this biased ensemble , choosing positive @xmath1 enhances the probability weights of trajectories with a large lyapunov indicator @xmath103 , favors via eq . the sampling of reactive paths passing over saddles and unstable directions of the potential energy landscape .
on the contrary , choosing negative @xmath1 mainly restricts the sampling of non reactive or regular trajectories within stable basins . the distribution @xmath120 is approximated by a markov chain of _ m _ steps constructed by importance sampling , by means of the metropolis algorithm .
the sampling is done in the following way : at markov chain step @xmath121 , starting from the current path @xmath122 , a trial path @xmath123 is generated with probability @xmath124 .
then , the trial path is accepted with a probability @xmath125 and added to the markov chain as @xmath126 ; otherwise , if the trial path is rejected , @xmath127 . to ensure the convergence of the markov chain towards the equilibrium distribution @xmath128
, we impose that the probability @xmath129 $ ] to transit from a path @xmath104 to a different path @xmath130 satisfies the detailed balance equation @xmath131\pi\left[\mathbf{z}\rightarrow\mathbf{z}'\right]=\mathcal{p}^{\alpha}_a\left[\mathbf{z}'\right]\pi\left[\mathbf{z}'\rightarrow\mathbf{z}\right]\ ] ] taking account of the generating and acceptance probabilities @xmath124 and @xmath125 , the transition probability @xmath132 reads @xmath133=\sum_{\tilde{\mathbf{z}}}p_{gen}\left[\mathbf{z}\rightarrow\tilde{\mathbf{z}}\right]\left\{\delta(\tilde{\mathbf{z}}-\mathbf{z}')p_{acc}\left[\mathbf{z}\rightarrow\tilde{\mathbf{z}}\right]+\delta(\tilde{\mathbf{z}}-\mathbf{z})(1-p_{acc}\left[\mathbf{z}\rightarrow\tilde{\mathbf{z}}\right])\right\}\ ] ] where @xmath134 is the delta distribution and @xmath130 is either the old path @xmath104 or the proposed path @xmath135 .
the acceptance probability can be constructed from eqs . and as the metropolis acceptance @xmath136=\min\left\ { 1,\frac{\mathcal{p}^{\alpha}_a\left[\tilde{\mathbf{z}}\right]p_{gen}\left[\tilde{\mathbf{z}}\rightarrow\mathbf{z}\right]}{\mathcal{p}^{\alpha}_a\left[\mathbf{z}\right]p_{gen}\left[\mathbf{z}\rightarrow\tilde{\mathbf{z}}\right]}\right\}.\label{eq:11}\ ] ] that is widely used in numerical simulations , as it has the main advantage of maximizing @xmath125 .
the standard shooting algorithm for deterministic dynamics is obtained by ( i ) selecting a state @xmath137 of the current trajectory , ( ii ) perturbing the momenta of each particle , and then generating from this selected state two segments , one backward of duration @xmath138 and the other one forward of duration @xmath139 , in order to get a trial trajectory @xmath135 of same duration , ( iii ) accepting or rejecting the new trajectory @xmath135 . for deterministic dynamics
, the total energy is constant .
the perturbation step ( ii ) is done with the algorithm proposed by stoltz @xcite , where trial momenta @xmath140 are obtained from old momenta @xmath141 as @xmath142 where @xmath143 is a tunable parameter and @xmath144 is drawn from a gaussian distribution of variance @xmath145 .
the probability of having trial momenta @xmath140 from old momenta @xmath141 with the stoltz proposal for shooting @xcite is written as @xmath146 and ensures the detailed balance condition @xmath147 hence the probability flux between the current and perturbed momenta is balanced and the hamiltonian distribution is preserved .
the probability @xmath148 $ ] of obtaining the shooting point @xmath149 for the trial trajectory from state @xmath137 selected in the current trajectory reads @xmath150=p\left(\mathbf{p}_{t'}\rightarrow\tilde{\mathbf{p}}_{t'}\right)\ ] ] as only momenta are perturbed at the shooting point , while positions are left unchanged . inserting eq . in eq . and
neglecting numerical approximations in the integration from @xmath151 to @xmath152 ( that indeed give in computations @xmath153 ) we have the metropolis acceptance rule @xmath154=\min\left\ { 1,\frac{\exp\left\ { \alpha l(\tilde{\mathbf{z}})-\beta \mathcal h(\tilde{\mathbf{x}}_0)\right\}\varphi_{a}^{\alpha}(\tilde{\mathbf{x}}_{0 } ) p_{gen}[\tilde{\mathbf{x}}_{t'}\rightarrow \mathbf{x}_{t ' } ] } { \exp\left\ { \alpha l(\mathbf{z})-\beta \mathcal h(\mathbf{x}_0)\right\}\varphi_{a}^{\alpha}(\mathbf{x}_{0})p_{gen}[\mathbf{x}_{t'}\rightarrow \tilde{\mathbf{x}}_{t ' } ] } \right\}\ ] ] where terms deriving from the forward and backward generation of the current and trial path have cancelled because of the unit phase space compressibility of the newtonian dynamics ( see ref .
@xcite for details ) . using in eq .
the property of the stoltz proposal ( eq . ) and neglecting numerical approximations in the integration from @xmath151 to @xmath152 , eq .
further simplifies in @xmath155=\min \left\ { 1,\exp\left\ { \alpha l(\tilde{\mathbf{z}})-\alpha l(\mathbf{z})\right\}\left[\varphi_{a}^{\alpha}(\tilde{\mathbf{x}}_{0})/\varphi_{a}^{\alpha}(\mathbf{x}_{0})\right ] \right\}.\ ] ] the second monte carlo move in trajectory space is based on the shifting algorithm , supplemented with a waste - recycling estimator .
@xcite @xmath52 trial trajectories @xmath156 are constructed from @xmath100 as follows : the duration of the trajectory @xmath157 is doubled selecting a random duration @xmath158 and integrating two segments backward and forward , starting from @xmath7 and @xmath159 , along @xmath160 and @xmath161 time steps , respectively . adding these segments to the current trajectory , one obtains a `` buffer '' trajectory @xmath162 of total duration @xmath163 , containing @xmath164 possible trial paths .
the conditional probability of obtaining the `` buffer '' trajectory @xmath165 starting from the current trajectory @xmath104 is indicated as @xmath166 .
the probability weight of each trial path @xmath156 is then@xmath167 if a constraining function @xmath168 linking the initial state @xmath169 of each trial path @xmath170 trajectories to state a is used , as in eq . .
j _ runs over the _ @xmath164 _
possible paths on the `` buffer '' trajectory .
we introduce an action @xmath171 so as to rewrite the trajectory probability in the biased ensemble ( eq . ) as @xmath172 we now define the selecting probability @xmath173 of selecting a trial trajectory @xmath170 from the buffer trajectory @xmath165 as @xmath174 where we introduced the rosenbluth factor @xmath175.\ ] ] resorting to bayes theorem , @xmath176 can be written as the posterior likelihood probability @xcite of having @xmath170 given the `` buffer '' trajectory @xmath165 : @xmath177 where @xmath178 is the conditional probability of constructing a `` buffer '' path @xmath165 from the trial path , and because of the deterministic dynamics @xmath179 .
@xmath180 is the marginal probability associated with the buffer path : comparing eq . with eq .
, we see that @xmath181 let us now consider a monte carlo move between two paths both contained in the buffer trajectory @xmath182 , i.e. from the current path @xmath104 to @xmath170 , whose associated transition probability @xmath183 $ ] , as in eq . , obeys a detailed balance with respect to the prior distribution @xmath120 : @xmath184\pi\left[\mathbf{z}\rightarrow\tilde{\mathbf{z}}_{j}\right]=\mathcal{p}^{\alpha}_a\left[\tilde{\mathbf{z}}_{j}\right]\pi\left[\tilde{\mathbf{z}}_{j}\rightarrow\mathbf{z}\right].\ ] ] defining the transition probability @xmath185=p_{sel}(\tilde{\mathbf{z}}_{j}|\mathbf{\zeta})p_{cond}(\mathbf{\zeta}|{\mathbf{z}})\ ] ] the detailed balance in eq .
can be recasted as @xmath186 where @xmath187 and @xmath188 is the probability to transit from @xmath170 to @xmath104 . as a consequence ,
the shifting procedure leaves the probability distribution @xmath120 invariant . moreover , recalling that the deterministic dynamics entails @xmath189 , the detailed balance in eq .
further simplifies into @xmath190 we conclude this section by extending the detailed balance of eq . for trajectories
@xmath104 and @xmath170 belonging to two _ different _ buffer paths @xmath182 and @xmath191 respectively . writing as in eq .
expressions for @xmath192 and @xmath193 , and using these results in eq .
we obtain @xmath194=p_{marg}^{\alpha}(\tilde{\zeta})p_{sel}(\tilde{\mathbf{z}}_{j}|\tilde{\mathbf{\zeta}})\pi\left[\tilde{\mathbf{z}}_{j}\rightarrow\mathbf{z}\right]\ ] ] eq .
is indeed a detailed balance condition for the alternate shooting and shifting moves , and samples the buffer trajectories @xmath182 .
this implies that the distribution @xmath195 is invariant along the sampling algorithm .
@xmath195 is therefore suitable to be used as an input path probability weight required by the unbiasing algorithm mbar , that will be presented and used in the following sections .
here we briefly recall how reaction rates are computed in the tps framework .
the reactivity of the sampled paths is recovered by means of a time correlation function with respect to initial and final states of the paths : a trajectory is said to be reactive if it starts in the reactants a basin and ends in the products b basin .
the time correlation function @xcite is @xmath196 where brackets @xmath197 indicate averages taken over the equilibrium trajectory ensemble and the indicator function @xmath198 is defined as in eq . .
@xmath199 can be understood as the conditional probability of observing a trajectory of duration @xmath200 ending in state b , knowing that it started in state a @xcite , and approaches its asymptotic value exponentially as @xmath201 where @xmath202 is the equilibrium occupation probability of state b , and the parameter @xmath203 is the characteristic reaction time of the system , given by the forward and backward reaction constants @xmath204 and @xmath205 , respectively .
note that the basic assumption required to compute reaction rate constants of rare events from the correlation function @xmath199 is the presence of a well separated time scale for processes occurring between fast intra - funnel relaxation , having a typical time constant @xmath206 , and activated processes indicating passages _ between _ funnels , needing a much longer time scale , of the order of @xmath207 @xcite . for times in the intermediate time regime @xmath208 ,
the correlation function in eq . can indeed be expressed by its first order expansion @xmath209 where the detailed balance condition @xmath210 has been used to eliminate @xmath202 .
hence , the slope of @xmath199 for this intermediate time regime gives direct access to reaction rates .
the reactive probability flux flowing from state a towards b per unit time , defined by @xmath211 , displays a plateau corresponding to the forward phenomenological reaction constant @xmath204 .
@xcite let us point out that these results , obtained with a macroscopic flux over population probability approach , can be recovered by a bennett - chandler formalism @xcite , based on microscopic quantities ( positions and momenta , see ref .
finally , we recall that the phenomenological forward rate constant @xmath204 computed from eq .
is related to the rate @xmath212 given by the transition state theory ( tst ) @xcite @xmath213 where @xmath214 is the _ transmission factor _ , and @xmath212 reads @xmath215 with @xmath216 the height of the free energy barrier separating basins a and b , @xmath25 the planck constant and @xmath217 the inverse temperature .
the transmission factor @xmath218 is always lower than one , and is introduced to take into account trajectories started in basin a that reach the saddle point but fall back to state a , instead of ending in state b : these occurrences are called recrossings events .
the transmission coefficient usually reaches a steady value , depending on the temperature and the reaction coordinate chosen to localize the barrier . at this plateau value of @xmath219 , and
the reactive flux corresponds to @xmath220 , always lower than the tst value . in numerical experiments ,
the computation of reaction constants by direct evaluation of @xmath199 at times longer than the intrafunnel relaxation time @xmath206 means performing very long molecular dynamics trajectories . to estimate @xmath199 in the tps framework
computing relatively short trajectories , one resorts to a factorization of the correlation function in ( i ) a static quantity related to @xmath212 , thus dependent on the free - energy barrier and calculated through an umbrella - sampling technique , and ( ii ) a dynamic factor related to @xmath214 given by the time derivative of the probability of reaching basin b at times shorter than the whole trajectory length ( see for instance refs .
@xcite ) we propose herein a variant strategy : reaction constants will be calculated directly by averaging indicator functions on short trajectories , as in eq .
, once the fraction of reactive paths is significantly enhanced by introducing an appropriate bias favoring reactions between basins a and b. the correlation function in eq . can be intended as an average over all performed trajectories - i.e. over all successive steps _ m _ of the markov chain - of the _ reactivity _
@xmath221 , defined as @xmath222 for an unbiased tps simulation , the probability @xmath120 of eq . with @xmath223
is sampled : we have @xmath224 , where brackets @xmath225 correspond to averages over a canonical trajectory ensemble , and the trajectory distribution of eq .
ensures @xmath226 because of @xmath109 . in a context of biased tps ,
averages on markov chains are taken on the biased trajectory distribution , hence we denote biased averages of the correlation function as @xmath227 , where index @xmath1 accounts for the current bias . herein , index @xmath1 will indicate all observables obtained from a biased distribution , where @xmath223 stands for the equilibrium canonical ensemble average .
we estimate the correlation function @xmath228 using an estimator denoted by @xmath229 , which consists in taking the average over the markov chain of lenght @xmath230 as @xmath231=\frac{1}{m}\sum_{m=1}^{m}\mathcal{a}^{m}_{\alpha}\ ] ] @xmath232 denotes a reactivity value coming from a biased trajectory @xmath233 distributed according to @xmath120 .
estimates given by @xmath234 are however not optimal @xcite .
a _ waste - recycling _ ( wr ) estimator @xcite is associated to the multiple proposal sampler for the shifting move of sec .
[ waste - recycling ] to obtain more accurate estimates .
waste recycling consists in including information about all possible proposals contained in the buffer trajectory @xmath182 . the reactivity at markov chain step _
m _ in a given ensemble @xmath1 has to be first averaged over the @xmath164 trajectories contained in the buffer path @xmath235 as @xmath236 } { \mathcal{r}^{m}_{\alpha}}=\sum_{j=1}^{\mathcal n}\mathcal{a}^{m}_{\alpha , j}\varphi_{a}^{\alpha}(\mathbf{x}^m_{0,j})\exp\left[-s^{m}_{\alpha , j}+s^{m}_{\alpha}\right]\ ] ] where we write @xmath237 and define the rosenbluth factor @xmath238}\ ] ] as proportional to the marginal probability @xmath239 of eq . , associated to the `` buffer '' trajectory @xmath240 corresponding to markov chain step @xmath121 .
correlation functions for reactive paths are therefore estimated in a way similar to eq .
, with the wr estimator @xmath241=\frac{\sum_{m=1}^{m}\sum_{j=1}^{\mathcal n}\mathcal{a}^{m}_{\alpha , j}\varphi_{a}^{\alpha}(\mathbf{x}^m_{0,j})\exp\left [ -s^{m}_{\alpha , j}+s^{m}_{\alpha}\right]}{\sum_{m=1}^{m}\sum_{j=1}^{\mathcal n}\varphi_{a}^{\alpha}(\mathbf{x}^m_{0,j})\exp\left [ -s^{m}_{\alpha , j}+s^{m}_{\alpha}\right]}=\frac{1}{m}\sum_{n=1}^{m}\overline{\mathcal{a}}^{m}_{\alpha}\ ] ] and again @xmath242 $ ] . the calculation of rate constants @xmath243 for the given @xmath1-ensemble follows from eq . .
a suitable unbiasing algorithm is needed in order to recover canonical ensemble correlation functions from reactivity values witnessed in a lyapunov biased path ensemble .
the canonical equilibrium values of @xmath244 can in principle be obtained by estimating reactivities @xmath245 , computed in any @xmath1-biased path ensemble , resorting to an adequate unbiasing algorithm .
we define an unbiasing wr estimator @xmath246 , where the left subscript @xmath247 indicates the ensembles in which we are interested in measuring averages , while the right subscript @xmath1 refers to the ensemble that our lyapunov biased tps will effectively sample .
equilibrium values for @xmath244 ( @xmath248 ) are retrieved from reactivities computed in any @xmath1-biased ensemble as @xmath249=\frac{\sum_{n=1}^{m}\sum_{j=1}^{\mathcal n}\mathcal{a}^{m}_{\alpha}\varphi_{a}^{\alpha}(\mathbf{x}^m_{0,j}){\varphi_{a}^{\alpha}}^{-1}(\mathbf{x}^m_{0,j})\exp\left [ -s^{m}_{\alpha , j}+s^m_{\alpha}+s^{m}_{\alpha , j}\right ] } { \sum_{n=1}^{m}\sum_{j=1}^{\mathcal n}\varphi_{a}^{\alpha}(\mathbf{x}^m_{0,j}){\varphi_{a}^{\alpha}}^{-1}(\mathbf{x}^m_{0,j})\exp\left [ -s^{m}_{\alpha , j}+s^m_{\alpha}+s^{m}_{\alpha , j}\right]}\ ] ] unbiasing the sampling consists of correcting for the bias @xmath250 $ ] .
however , the variance associated to the unbiasing estimator in eq . would be too large to consider estimates reliable @xcite .
this well - known fact results from the lack of overlap between the sampled and measured distributions .
we therefore carry out a series of simulations for a set of @xmath1 values ranging from 0 to a maximum value @xmath251 so to ensure overlap between successive sampled distributions .
our choice is then to use the multistate bennett acceptance ratio ( mbar ) instead of the estimator of eq . .
mbar is a method elaborated by shirt and chodera @xcite , which aims at minimizing the statistical variance associated to the estimates . following these authors
, we briefly expose the principles of this procedure in a context of biased path ensembles . to use the mbar method in the waste - recycling framework of sec .
[ rarecostbiased ] , we take as probability weights corresponding to a given @xmath1 ensemble at each markov chain step @xmath121 the marginal probability @xmath252 of eq . .
this is possible because , thanks to the detailed balance of eq .
, @xmath195 is preserved .
once one knows weights @xmath253}$ ] related ( via eqs . and )
to @xmath254 for each @xmath1-ensemble , reactivity averages @xmath255 for every ensemble @xmath256 can be computed resorting to the importance sampling identity @xmath257}\right\rangle _ { \alpha}}{\left\langle \mathcal{a}\exp{\left[-s_{\alpha}\right]}\right\rangle _ { \alpha'}}=\frac{z_{\alpha'}}{z_{\alpha}}\label{eq:10}\ ] ] where we used the partition functions @xmath258 } $ ] . for a set of _ k _ different values of the bias @xmath1 , a set ( namely , a markov chain ) of @xmath259 buffer trajectories are sampled for each bias value .
an estimate of @xmath260 is given by the mbar estimator @xmath261 for the waste - recycling averaged reactivity @xmath262 of eq .
as @xmath263=\sum_{m=1}^{\mathcal{m}_{\alpha } } w_{m,\alpha } \overline{\mathcal a}^m_{\alpha}\ ] ] where weights @xmath264 are given by the expression @xmath265}}{\sum_{k=1}^{k}\mathcal{m}_{k}\hat{z}_{k}^{-1}\exp{\left[-s_{\alpha}^m\right]}}.\ ] ] and @xmath266 are estimators for the partition functions @xmath267 with minimal asymptotic covariance ( see eq . and ref .
note that the denominator in eq .
indicates that each weight @xmath264 takes into account contributions from all other ensembles @xmath268 introducing also partition functions @xmath269}$ ] , the uncertainty can be estimated as @xmath270)\approx{\mathbb{k}_{\alpha}[\mathcal a]}^2 ( \rm{var}(\hat z_{\mathcal a_{\alpha}})+ \rm { var}(\hat{z}_{\alpha})-2 \rm { cov}(\hat z_{\mathcal a_{\alpha}},\hat{z}_{\alpha})).\ ] ] canonical equilibrium values of the correlation function @xmath244 and corresponding values of the reaction rate constants can be recovered once we consider estimates in eq . with @xmath223 .
we emphasize that this is by now the first application of the mbar unbiasing method based on marginal probabilities , able to estimate observables computed with a path sampling algorithm supplemented by waste - recycling .
numerical recipes to obtain these estimations have been furnished by j. chodera @xcite .
in order to test the lyapunov biased tps algorithm ( lytps ) , we first consider a lennard - jones 38 cluster .
this well - known benchmark system is often uysed to assess the efficiency of sampling algorithms .
it and has been widely explored in literature for its rich thermodynamic properties : its potential energy landscape presents two main basins : a deep and narrow funnel containing the global energy minimum , a face - centered cubic truncated octahedron structure ( fcc ) , and a separate , wider , funnel leading to a large number of icosahedral structures ( ico ) of slightly higher energies .
although the configuration with the lowest potential energy corresponds to the fcc one , the greater configurational entropy associated with a large number of local minima in the icosahedral funnel make this second configuration much more stable at higher temperatures . as temperature increases
, lj@xmath0 undergoes several structural transitions .
first , a solid - solid transition occurs at @xmath271 when the octahedral fcc structure gives place to the icosahedral ones .
secondly , above @xmath272 , the outer layer of the cluster melts , while the core remains of icosahedral structure .
@xcite the lennard - jones potential is given by the usual expression @xmath273\ ] ] where @xmath274 is the position of the @xmath275-th atom , @xmath276 is the distance between atoms @xmath275 and @xmath277 , @xmath143 is the pair well depth and @xmath278 is the equilibrium pair separation .
in addition , a trapping potential prevents the evaporation of clusters particles at finite temperature : if the distance between the position @xmath279 of a particle and the center of the trap @xmath280 exceeds @xmath281 , this particle feels a potential @xmath282 .
the bond - orientational order parameters @xmath283 defined as @xcite @xmath284 where the @xmath285 are spherical harmonics and @xmath286 are the polar and azimuthal angles of a vectors corresponding to bonds between the @xmath287 pairs of atoms .
the @xmath288 are traditionally used to identify different crystalline symmetries .
the parameter @xmath289 is often used to distinguish between the icosahedral and cubic structures , for which it has values around 0.02 and 0.18 respectively .
@xcite monte carlo sampling fails to equilibrate the two funnels , and global optimization methods are unable to find its global energy minimum @xcite .
hence , several elaborated algorithm have been employed in the past to study the thermodynamic equilibrium of this system , such as parallel tempering , @xcite basin - sampling techniques , @xcite wang - landau approaches @xcite or path - sampling methods .
@xcite standard transition path sampling @xcite and discrete path sampling @xcite ( dps ) have been already used to study transitions between the two funnels of lj@xmath0 . however , in the case of tps the large number of metastable states separating the two main basins prevented the traditional shooting and shifting algorithm to identify reactive paths , despite previous success for smaller lj clusters .
@xcite authors had to resort to a two - ended approach linking the two minima to find trajectories with the same energy of those found by dps approach .
@xcite the main drawbacks of this tps method were a lack of ergodicity and a very large computational cost .
conversely , dps has been more succesfull in this task .
this method uses eigenvector following and graph transformation @xcite to compute the overall transition rate between two regions of phase space . to the best of our knowledge ,
this is by now the most successful approach to computing reaction rates in lj@xmath0 .
@xcite in particular , reaction rate constants for transitions between the two solid structures have been computed using dps @xcite at different temperatures .
we use here the lyapunov biased tps algorithm to investigate structural transitions in lj@xmath0 for temperatures above and below the solid - solid transition temperature @xmath290 , spanning a temperature range from @xmath291 to @xmath292 .
our simulations required about @xmath293 hours of cpu time to observe reactive trajectories between the two main funnels .
reaction constants , computed with the method exposed in sec . [ sec : reaction - rate - constants - calculation ] , can be compared to values obtained with the discrete path sampling approach .
@xcite in order to thermalize the system at a given temperature , langevin dynamics are run for @xmath294 time steps , using a friction parameter @xmath295 .
the last configuration is used as the starting point of the first deterministic trajectory of the tps simulations . at each new temperature
, a new preliminary langevin dynamics is performed . in simulations , trajectories consist of @xmath296 time steps , each step of duration @xmath297 .
deterministic trajectories are obtained with the verlet algorithm @xcite , and then selected following the lyapunov biased tps algorithm described in sec .
[ sec : transition - path - sampling ] .
for each temperature , 25 different biased path distributions are sampled , for values of the control parameter @xmath1 ranging linearly from @xmath298 to @xmath299 , in order to obtain reactive paths for @xmath299 and ensure a sufficient overlap between distributions sampled for different @xmath1 values .
the unbiased distribution corresponding to @xmath223 has been simulated with tps as well .
a markov sequence of @xmath300 biased tps shooting and shifting moves is performed , in order to ensure an ergodic sampling .
values for the control parameters are chosen after observing the magnitude of the lyapunov indicators @xmath103 for few trajectories , and the difference between lyapunov indicators @xmath103 for current and trial trajectories in the shooting step ( eq . ) , in order to have an acceptance ratio for the shooting move not below @xmath301 , see fig .
[ fig : acceptance ] .
the choice of the trajectory length @xmath48 depends not on the lyapunov bias , but on the necessity of having long enough trajectories to link the two funnels .
@xmath48 needs also to be long enough to recover an appropriate statistics to compute reaction contants from correlation functions ( see below ) .
the use of the stoltz algorithm ( eq . ) in the shooting moves ensures that the energy distribution imposed by the preliminary md is maintained along the simulation .
the value for @xmath143 in eq . is taken as @xmath302 , so to have decorrelated sampled paths and ensure a sufficient acceptance ratio , see fig .
[ fig : accept_stoltz ] . over a markov chain of @xmath294 trajectories ,
starting from the fcc basin at @xmath303 , as a function of the control parameter @xmath1 used in the simulations , for different trajectory lengths @xmath48 . increasing @xmath1 increases the mean lyapunov indicator and enables trajectories to explore barriers and transition states .
average lyapunov indicators are almost independent of the trajectory length @xmath48 . *
bottom * : acceptance ratio , given by eq .
for the same simulations . ] in the stoltz algorithm ,
eq . , for a markov chain of @xmath294 trajectories , starting from the fcc basin at @xmath303 , with trajectory lengths @xmath304 and a control parameter set to @xmath305 . ]
we focus on the octahedral to icosahedral ( fcc - ico ) transition at temperatures from @xmath291 to @xmath292 : observing this passage using a direct md or a standard tps would require a considerable amount of cpu time ( about @xmath306h , see ref .
@xcite ) as the fcc configuration is at low temperatures the most stable one , so that the system rarely escapes from the fcc basin .
in contrast , with our biased tps technique we were able to observe at @xmath303 the first fcc - ico reactive trajectories after about @xmath307 markov chain steps . to ensure that reactive paths start in the stable fcc state , we include in the path probability weight the constraining function ( see eq . )
@xmath308 assigned to the starting state @xmath309 of the path , function of the bond order parameter @xmath289 and centered on the value @xmath310 .
we set @xmath311 , a sufficiently small stiffness that lets the trajectory starting point span the whole fcc basin .
the function @xmath312 keeps the beginning of the trajectories inside the fcc funnel , thus counterbalancing the effect of the local lyapunov bias , that would pull trajectories on barriers .
we present in fig .
[ fig : histo_fc12 ] histograms for the first and the last point of the trajectories , for different values of the control parameter @xmath1 at the fcc - ico cohexistence temperature @xmath303 . as
@xmath1 values increase , trajectories explore regions that are increasingly distant from the initial fcc basin , and some of them eventually cross the transition region and reach the ico basin . for trajectories starting from the fcc basin at @xmath303 for different values of the control parameter @xmath1 , averaged on a markov chain of 5000 steps .
the restraining function of eq . mantains the initial states of the trajectories in the fcc funnel for all @xmath1 values .
* bottom * : same histogram , for the final position @xmath313 .
trajectories sampled with large @xmath1 values escape the fcc funnel more often .
their final states are distributed over the whole fcc - ico range . ]
once reaction paths have been identified , the computation of the inter - funnel reaction constant by the correlation function via eq . of sec .
[ sec : reaction - rate - constants - calculation ] is possible if reactants and products basins are adjacent , i.e. if there is no intermediate state between them .
@xcite however , this hypothesis is not valid for the fcc - ico transition : several results reported in the literature @xcite show that reactive paths linking fcc and ico states pass through many short - lived metastable basins , separated by barriers of different heights , not belonging to the two main funnels .
these metastable states and transition regions have also been observed in a previous work using the transition current sampling method .
@xcite such a feature has been confirmed as well by an attentive analysis of our trajectories . among all the intermediate metastable states ,
we emphasize the presence of a basin related to a faulted fcc configuration , indicated with d in the following , having a bond order parameter value around @xmath314 , already acknowledged in precedent studies .
@xcite this basin has a rather important occupation probability if compared to other metastable states , and is visited by every reaction path linking fcc to ico state .
moreover , this metastable state is indeed structurally related to the fcc basin , and the barrier separating the d structure from fcc is lower than the one separating the former from ico state . as a result , several recrossing events of trajectories starting in fcc , visiting the d state and then going back to fcc , can be observed .
hence , in order to correctly reconstruct the fcc to ico transition paths , we have to take into account this intermediate metabasin .
we therefore split the fcc - ico passage in two steps : the first part is given by the passage from the fcc basin to the d basin corresponding to @xmath314 .
the second part is then given by trajectories starting from the d basin , and ending up in the ico funnel . to obtain this second part of fcc - ico reactive paths ,
we constrain the first point of the trajectories to start in the d basin , using a constraining function given by an indicator on the bond - order parameter value @xmath315 : @xmath316 for trajectories starting from the d configuration metabasin located at @xmath317 for different values of the control parameter @xmath1 , averaged over a markov chain of 5000 steps . *
bottom * : same histogram , for the final position @xmath313 .
trajectories end up in both the fcc or the ico funnel . ] in fig .
[ fig : histo_dc13 ] we present histograms for the distribution of the beginning and the end point of trajectories constrained with the indicator function of eq .
, at a temperature @xmath317 slightly above the solid - solid cohexistence .
paths sampled with the unbiased distribution @xmath223 completely remain in the `` window '' given by @xmath318 . on the contrary ,
trajectories weighted with a lyapunov bias tend to leave the metabasin : their starting points @xmath309 tend to accumulate on the borders of the region defined by the indicator function in eq .
, while the end points @xmath313 fall both in the fcc and the ico funnels . in simulations performed at lower temperatures ,
this reconstruction of the second part of the fcc - ico reactive path with trajectories starting from the d state is more difficult .
indeed , histograms for trajectories of the same length at temperatures lower than the solid - solid transition @xmath303 show that an important fraction of the sampled trajectories fall from the d state directly to the fcc basin , while a few trajectories end in the ico state .
this is attributed to the heights of the barriers separating the metastable d structure from either the stable ico or stable fcc structures , the latter barrier being lower than the former one .
the total reaction constants for the two - step fcc - ico transition , assuming a steady occupation probability for the intermediate d state , is derived in appendix [ react_cost_interm ] and reads @xmath319 where subscripts @xmath320 , @xmath321 and @xmath322 refers to fcc , d and ico states respectively .
the same steady state approximation is assumed for all intermediate metastable states in discrete path sampling studies .
@xcite reaction rates @xmath323 and @xmath324 involve transitions between states separated by high free energy barriers , @xcite thus the hypothesis of time scale separation required by the reaction rate theory is still valid , and reaction constants can be computed using the method exposed in sec . [
sec : reaction - rate - constants - calculation ] .
reactive paths between fcc and d basins , and between d and ico basins , are computed as reported above ( sec .
[ ljreactivepaths ] ) .
on the contrary , the d to fcc reaction rate @xmath325 can not be computed by lytps , because the requirement of a time scale separation is no longer valid , the barrier separating this two states being too low .
it is therefore computed by direct md simulation . the reactivity @xmath326 ( eq . ) for each trajectory
is evaluated in simulations distinguishing the three basins fcc , ico and d whose ranges of bond - order parameter @xmath327 value , that is @xmath328 , @xmath329 and @xmath330 , respectively .
data harvested during lytps runs are unbiased using mbar . in fig .
[ fig : corr_fc13 ] and [ fig : corr_dc13 ] , two examples of population correlation functions for the computation of reaction rate constants , unbiased with mbar , are reported .
note that reactivity values computed at short times are nearly zero , and do not contribute significatively to the correlation functions : in fact , these values are obtained from segments of trajectories too short to witness a complete transition between two states . in table
[ table ] , we report reaction rate constants values for the fcc to d structure ( @xmath323 ) transition , and fro the d structure to ico ( @xmath324 ) , that give , through eq . , a total fcc to ico ( @xmath331 ) rate in good agreement with values given by dps calculations @xcite . finally , an arrhenius plot comparing our results with the reaction constants proposed in ref .
@xcite is presented in fig .
[ fig : arrhenius_lj38 ] . .
error bars are computed via eq . and directly given by chodera s numerical recipe @xcite .
* right * : reaction constant for this same passage , obtained at times shorter than the mean first passage time .
the reactive flux @xmath332 reaches a plateau value , corresponding to @xmath323 , as explained in sec .
[ sec : reaction - rate - constants - calculation].,width=755 ] . *
bottom * : reaction constant for this same passage.,width=755 ] .table of reaction constants for the transitions from fcc to d structure , d to ico , and the total fcc to ico transition , indicated as @xmath323 , @xmath324 and @xmath331 respectively , at different temperatures .
values of @xmath331 are obtained using eq . and assuming the reaction constants @xmath325 as @xmath333 , @xmath334 , @xmath335 for @xmath291 , @xmath336 and @xmath303 respectively ( values obtained by langevin md ) , and unitary for @xmath337 . in the last column on the right , we report dps data from ref .
@xcite , computed in the harmonic approximation framework . [ cols="^,^,^,^,^",options="header " , ] ( lytps , red dots ) compared with data obtained from dps @xcite using an harmonic approximation and reported in ref .
@xcite ( ha , black line ) . ]
the second example of a thermally activated process studied using lytps is the migration of a single vacancy in @xmath1-iron crystal .
atomic interactions of the model system are described by an embedded atom model potential .
@xcite the crystal structure is body - centered cubic , and the initial unrelaxed cell contains 1023 atoms displayed on 1024 lattice sites , the vacant site corresponding to the vacancy .
the reaction coordinate used to represent the motion of the vacancy is the distance crossed by the moving atom that replaces the vacancy .
the free energy landscape for this system has been reconstructed in ref .
it presents two stable states , the first corresponding to the initial configuration , and the second one to the same configuration obtained by translating a neighboring atom into the vacancy site .
the first neighbor distance is @xmath338 , switching its initial position with the vacancy site .
there is a single free energy barrier separating these two states for temperatures above @xmath339 , while for lower temperatures an intermediate metastable state appears , corresponding to an intra - site position for the moving atom .
@xcite we performed lytps simulations with trajectories of different lengths ( see below ) , with time step @xmath340 .
a preliminary md simulation is done to equilibrate the system to the required temperature , with a friction parameter @xmath341 .
we explored temperatures ranging from @xmath342 to @xmath343 .
the lytps shooting and shifting moves are iterated to cosntruct markov chains consisting of @xmath344 trajectories . as for lj@xmath0 ,
the trajectory length and the values for the control parameter @xmath1 have been chosen in order to ensure an acceptance ratio of @xmath345 and an adequate ergodic sampling of the phase space . for temperatures above @xmath346
, the presence of a single `` smooth '' barrier separating the two metastable states makes this application simple enough : sampling of reactive trajectories is achieved using @xmath347 @xmath1 values from the unbiased simulation at @xmath223 up to @xmath348 , with trajectories of @xmath307 steps . for temperatures below @xmath349 , an ergodic sampling of trajectory space appears more difficult .
we therefore employed longer trajectories of @xmath350 time steps , as well as larger values of the control parameter , up to @xmath351 to allow the system to escape the initial basin .
reaction constants for the passage between the two stable states above @xmath349 are estimated from correlation functions unbiased with the mbar algorithm , via eq . .
for @xmath352 , the presence of an intermediate metastable basin has to be taken into account in the evaluation of reaction constants . as recalled in sec .
[ lj38 ] , the reaction rate expression obtained from eq . holds only for adjacent reactant and product basins . at low temperatures ,
it is therefore more appropriate to use our algorithm to evaluate the reaction constant for the passage from the initial state to the intermediate basin . from eq .
, reaction rate for the passage from one stable configuration to the other is simply half the reaction constant from one stable configuration to the intermediate one . indeed ,
reaction constants for transitions from the intermediate metastable state to either of the two stable states are equal , because of the symmetric shape of the potential surface .
@xcite in fig .
[ fig : arrhenius_vac ] we compare the reaction rates obtained with lytps , those computed inserting in the transition state theory ( tst ) expression ( eq . )
the free energy barriers reported in ref .
@xcite , and reaction constants obtained with a classical harmonic approximation ( ha ) . above the debye temperature ( @xmath353 ) , rates obtained with lytps fall between tst and harmonic approximation values .
to explain this point , we recall that reaction rates we estimate with the method exposed in sec . [
sec : reaction - rate - constants - calculation ] are derived from eq . ,
hence correspond to the _ phenomenological _ rate constants .
these values are therefore bounded from above by tst values , that overestimate reaction rates @xcite , as can be seen from eq . .
conversely , values obtained with the harmonic approximation neglect anharmonicity effects on the activation barrier , thus giving reaction rates that are lower then the phenomenological rate constants we compute .
our results are then in agreement with the reaction rate theory recalled in sec .
[ sec : reaction - rate - constants - calculation ] .
the method presented in this paper allows to compute reaction rate constants for inter funnel transitions in many - body systems .
the reaction rate values are evaluated using a path sampling algorithm biased with local lyapunov numbers .
this bias aims at enhancing an accelerate sampling of reactive paths , so as to reduce the lenght of markov chains and the amount of cpu time to observe activated processes .
we assess the performace of the method by observing reaction paths and evaluating equilibrium reaction rates for structural transitions in the lj@xmath0 system and for vacancy migration in an @xmath1-iron crystal . for both systems
, we were able to predict phenomenological rate constants , in very good agreement with data already given in the literature in the case of lj@xmath0 .
the lyapunov biased tps method presents several advantages , and incorporates features of different rare events simulation methods .
firstly , with respect to other importance sampling methods based on lyapunov weighted sampling @xcite , lyapunov biased tps has the main advantage of a simpler implementation .
this is due to the lyapunov indicator @xmath102 we propose in eq .
, that allows to quantify the divergence of hamiltonian trajectories by resorting to local lyapunov numbers .
these quantities can be easily calculated with the lanczos algorithm , that enables one to compute the largest eigenvalues of the jacobian matrix of the hamiltonian mapping with a limited computational cost . as recalled in sec .
, resorting to local lyapunov numbers to evaluate chaoticity of phase space trajectories does nt suffer from the computational drawbacks of other algorithms aimed at the same purpose , as rli or the tangent space method .
@xcite the implementation of shooting and shifting monte carlo moves in a lyapunov biased tps is less complicated , and computationally less expensive , than the algorithm proposed in ref .
@xcite with the use of rli , because we do not need to compute four trajectories to evaluate the divergence of a single path @xcite . secondly , this formulation for the lyapunov indicator is such that the bias applied to each path in order to enhance the fraction of reactive trajectories is clearly identified , contrarily to what is done in rather complex cloning algorithms like the one proposed in lyapunov weighted dynamics @xcite and transition current sampling @xcite .
hence , the use of standard unbiasing statistical tools to recover unbiased observables is possible and requires a small theoretical and computational effort .
furthermore , we consider the access to the evaluation of equilibrium transition rates as the most important aspect of lyapunov biased tps . on the computational point of view , the direct access to reaction rates without resorting to a distinct evaluation of the reaction barriers and the transmission factor , as usually done in standard tps technique @xcite , is a very advantageous feature . to unbias reaction constants computed in lyapunov biased ensembles we chose among other unbiasing algorithm , like wham @xcite or extended bridge sampling @xcite techniques , the mbar method @xcite .
mbar has proven to be computationally efficient and to give an adequate numerical precision in estimating reaction constants . moreover ,
this work is the first in which mbar is implemented exploiting the marginal probability derived from a waste recycling method .
this allows to include informations encoded in rejected proposal trajectories .
finally , lytps is implemented at a finite temperature , imposed to trajectories by the canonical distribution from which the path starting point is selected and maintained along the path thanks to the stoltz proposal for the shooting algorithm , see sec .
[ sec : transition - path - sampling ] . in parallel ,
the lyapunov indicator used as a bias to select reactive paths directly links the path sampling to the local conformation of the potential energy surface via the hessian matrix , thus giving to our method an intrinsic dependence on the potential energy landscape .
the coupling between a finite - temperature sampling and potential energy surface conformation is a noticeable improvement if compared to eigenvector - following methods , that are based on the shape of the potential energy surface , but usually operate at zero temperature .
lytps can be acknowledged as a finite , nonzero temperature version of the well - known eigenvector - following techniques , such as dimer , optim or art .
@xcite the advantages related to a finite temperature technique do not concern only the exploration of the energy landscape , but also the fact that the evaluation of physical observables like reaction rates takes into account temperature and anharmonicity effects .
indeed , the phenomenological reaction rate we computed ( see sec . [
sec : reaction - rate - constants - calculation ] ) can be compared with experimental measures : lytps turns out to be a powerful tool to study problems like vacancy migration in iron . in this
context , reaction rates are usually estimated using only the _ potential energy _
barriers at 0 k and harmonic approximations : this give approximate results with respect to experimental data obtained at nonzero temperatures .
lytps reaction rates find an important application as input parameters for object monte carlo codes aimed at numerically reproducing resistivity recovery experiments .
we conclude observing that this method can be furhter developed using a parallel tempering technique , or implemented for the computation of reaction rates in more complex condensed matter systems , and can find interesting applications in a wide class of research fields , spanning from molecular biophysics to physical metallurgy , where the numerical determination of reaction rates has important consequences for experimental applications .
i warmly thank manuel athenes and mihai - cosmin marinica for help and fruitful discussions .
the time evolution of the occupation probabilities @xmath354 , @xmath355 and @xmath356 of the fcc , ico and faulted states respectively , read @xmath357 where subscripts @xmath320 , @xmath321 and @xmath322 refers to fcc , faulted and ico states respectively , and @xmath204 indicates the generic transition rate from state a to state b. note that the possibility of a direct transition from fcc to ico states without passing by the intermediate basin has been neglected .
the stationary approximation for the metastable state imposes @xmath358 , thus the above system can be recasted in the form @xmath359 and the reaction constant between fcc and ico states can be calculated as @xmath360 | we propose an efficient method to compute reaction rate constants of thermally activated processes occurring in many - body systems at finite temperature .
the method consists in two steps : first , trajectories are sampled using a transition path sampling ( tps ) algorithm supplemented with a local lyapunov bias favoring diverging trajectories .
this enhances the probability of observing rare reactive trajectories between stable states during relatively short simulations .
secondly , reaction constants are eventually estimated from the unbiased fraction of reactive trajectories , yielded by an appropriate statistical data analysis tool , the multistate bennett acceptance ratio ( mbar ) package . in order to test our algorithm ,
we compute reaction constants for structural transitions in lj@xmath0 , a well studied lennard - jones cluster , comparing our results to values previously reported in the literature .
additionally , we apply our method to the calculation of reaction rates for the migration of vacancies in an @xmath1-iron crystal , for temperatures ranging from 300 k to 850 k. vacancy diffusion rates associated with activation barriers at finite temperature are then evaluated , and shown to be substantially different from values obtained using the standard harmonic approximation . |
let @xmath0 be a bounded polygonal or polyhedral domain with lipschitz boundary @xmath1 in @xmath2 .
consider the following fourth - order singularly perturbed elliptic equation @xmath3 with boundary conditions @xmath4 or @xmath5 where @xmath6 , @xmath7 is the standard laplace operator , and @xmath8 denotes the outer normal derivative on @xmath9 . in two dimensional cases ,
the boundary value problems ( [ pde1])-([boundar condition1 ] ) and ( [ pde1])-([boundary condition2 ] ) arise in the context of linear elasticity of thin bucking plate with @xmath10 representing the displacement of the plate .
the dimensionless positive parameter @xmath11 , assumed to be small ( i.e. , @xmath12 ) , is defined by @xmath13 where , @xmath14 is the thickness of the plate , @xmath15 is the young modulus of the elastic material , @xmath16 is the poisson ratio , @xmath17 is the characteristic diameter of the plate , and @xmath18 is the absolute value of the density of the isotropic stretching force applied at the end of the plate @xcite . in three dimensions , problems
( [ pde1])-([boundar condition1 ] ) and ( [ pde1])-([boundary condition2 ] ) can be a gross simplification of the stationary cahn - hilliard equations with @xmath11 being the length of the transition region of phase separation .
conforming , nonconforming , and mixed finite element methods for fourth order problem have been extensively studied @xcite . however , its _ a posteriori _
error estimation is a much less explored topic . even for the kirchhoff plate bending problem , the finite element _ a posteriori _ error analysis is still in its infancy . in 2007 , beir@xmath19o et al .
@xcite developed an estimator for the morley element approximation using the standard technique for nonconforming element .
later , hu et al .
@xcite improved the methods of @xcite by dropping two edge jump terms in both the energy norm of the error and the estimator , and by dropping the normal component in the estimators of @xcite .
therefore , a naive extension of the estimators in @xcite to the current problem may probably not be robust in the parameter @xmath11 .
designing robust _ a posteriori _ estimators is challenging , especially for singularly perturbed problems , since constants occurring in estimators usually depend on the small perturbation parameter @xmath11 .
this motivates us to think about the question : what method and norm are suitable for the singularly perturbed fourth - order elliptic problem ? in the literature , @xmath20 is a widely used measure for the primal weak formulation .
we recall _ a priori _ estimates in @xcite for boundary condition ( [ boundary condition2 ] ) and convex domain @xmath0 : @xmath21 hereafter , we use @xmath22 to denote a generic constant independent of @xmath11 with different value at different occurrence .
this leads to @xmath23 multiply both sides of ( [ pde1 ] ) by @xmath10 , and then integrate over @xmath0 . using integration by parts and boundary condition ( [ boundar condition1 ] ) or ( [ boundary condition2 ] ) , we have from the poincar inequality that @xmath24 as a consequence , @xmath25 .
this suggests that the two components of @xmath26 are unbalanced with respect to @xmath11 if @xmath6 .
furthermore , if we set @xmath27 , then problem ( [ pde1 ] ) is written as @xmath28 note that @xmath29 has boundary layer , but @xmath10 usually does not have one . thus , @xmath30 approaches to @xmath31 as @xmath32 , which fails to describe the layer of @xmath29 .
an observation of the two decoupled equations @xmath33 and @xmath34 suggests that the two measures @xmath35 and @xmath36 can portray the layer of @xmath29 and the first and second derivatives of @xmath10 . from ( [ maysp ] ) , we have @xmath37 notice that @xmath38 if @xmath6 , then @xmath39 and @xmath40 are balanced with respect to @xmath11 for the boundary condition ( [ boundary condition2 ] ) .
these inspire us to think about the mixed finite element method for the problem ( [ pde1 ] ) and the two aforementioned measures .
however , the mixed finite element method for the problem ( [ pde1 ] ) is a much less explored topic , since there exist some special problems such as the fourth order problem , where attempts at using the results of brezzi and babuka were not entirely successful since not all of the stability conditions were satisfied , cf . @xcite and the reference therein . to overcome this difficulty , falk et al . developed abstract results
from which optimal error estimates for these ( biharmonic equation ) and other problems could be derived ( @xcite ) .
however , it is not easy to extend the results of @xcite to the problem ( [ pde1 ] ) , because of the existence of an extra term and the singular perturbation parameter @xmath11 .
recently , for a fourth order reaction diffusion equation , the error estimates of its mixed finite element method was derived in @xcite .
we refer to @xcite about the _ a posteriori _ estimation of ciarlet - raviart methods for the biharmonic equation . in this work ,
our goal is to develop robust residual - type _ a posteriori estimators for a _ mixed finite element method for the problem ( [ pde1 ] ) in the two aforementioned measures .
the main difficulty lies in the fact that the boundary condition ( [ boundary condition2 ] ) does not include any information on the immediate variable @xmath29 . in order to overcome this difficulty
, we develop a novel technique to analyze residual - based _ a posteriori _ error estimator .
the key idea is to replace a function @xmath41 ( such that @xmath42 ) without boundary restriction by a function @xmath43 with boundary restriction , which catches at least @xmath44 times " of @xmath45 in the @xmath11-weighted energy norm ( see lemma 3.3 below ) . combining this novel design with standard tools , we develop uniformly robust residual - type _ a posteriori _ estimators with respect to the singularly perturbed parameter @xmath11 in the two aforementioned measures .
we refer to the reference @xcite on balanced norm for mixed formulation for singularly perturbed reaction - diffusion problems .
the rest of this paper is organized as follows : in section 2 , we introduce mixed weak formulations and some notations , and prove an equivalent relation between the primal weak solution and the weak solution determined by its mixed formulation .
some preliminary results are provided in section 3 .
residual - type _ a posteriori _ estimators are developed and proven to be reliable in section 4 .
an efficient lower bound is proved in section 5 . in section 6 ,
numerical tests are provided to support our theory .
setting @xmath27 , and employing the boundary condition ( [ boundar condition1 ] ) , we attain the ciarlet - raviart mixed problem @xmath46 : @xmath47 similarly , using the boundary condition ( [ boundary condition2 ] ) , we arrive at the ciarlet - raviart mixed formulation @xmath48 : @xmath49 for any bounded open subset @xmath50 of @xmath0 with lipschitz boundary @xmath44 , let @xmath51 and @xmath52 be the standard lebesgue and sobolev spaces equipped with standard norms @xmath53 and @xmath54 , @xmath55 ( see @xcite for details ) .
note that @xmath56 .
we denote @xmath57 the semi - norm in @xmath52 . similarly , denote @xmath58 and @xmath59 the @xmath60 inner products on @xmath44 and @xmath50 , respectively .
we shall omit the symbol @xmath0 in the notations above if @xmath61 .
the weak formulation of problem @xmath46 reads : find @xmath62 such that @xmath63 the weak formulation of the problem @xmath48 reads : find @xmath64 such that @xmath65 note that , by the lax - milgram lemma , both systems ( [ discrete mixed formulation p1 ] ) and ( [ discrete mixed formulation p2 ] ) have a unique solution .
in fact , by regularity theory for elliptic problems @xcite , if @xmath0 is convex and @xmath66 , then @xmath67 and @xmath68 . thus ( [ discrete mixed formulation p1 ] ) has solution , which is unique since its homogeneous system has only one solution satisfying @xmath69 .
similar conclusion can be drawn for the system ( [ discrete mixed formulation p2 ] ) .
it is well known that the primal weak formulation of ( [ pde1])-([boundar condition1 ] ) is : find @xmath70 such that @xmath71 and that the one of ( [ pde1])-([boundary condition2 ] ) is : find @xmath72 such that @xmath73 the classical results of pdes imply that ( [ may1 ] ) and ( [ may2 ] ) have unique solutions ( see @xcite ) .
a natural question is whether the @xmath10 determined by ( [ discrete mixed formulation p1 ] ) ( or ( [ discrete mixed formulation p2 ] ) ) is the solution of ( [ may1 ] ) ( or ( [ may2 ] ) ) . in @xcite , for biharmonic equation on a reentrant corners polygon , a counterexample is shown .
the following theorems answer this question .
[ mayy ] the solution @xmath74 of ( [ may1 ] ) and the @xmath10 determined by ( [ discrete mixed formulation p1 ] ) are identical if and only if @xmath75 .
the necessity is trivial . if the solution of ( [ discrete mixed formulation p1 ] ) is such that @xmath76 , then we have from the second equation @xmath77 notice that @xmath78 is dense in @xmath79 .
it follows that @xmath80 integration by parts yields @xmath81 which implies @xmath82 we obtain from ( [ may3 ] ) and the second equation of ( [ discrete mixed formulation p1 ] ) that @xmath83 in terms of ( [ may1 ] ) , we proved that @xmath10 is the solution of ( [ may1 ] ) .
the solution @xmath74 of ( [ may2 ] ) and the @xmath10 determined by ( [ discrete mixed formulation p2 ] ) are identical if and only if @xmath84 .
the necessity is trivial . from the second equation of ( [ discrete mixed formulation p2 ] ) , integration by parts , and variational principle , we know that the neumann boundary condition @xmath85 on @xmath9 is automatically satisfied . following the proof of theorem [ mayy ] , we know that if the solution @xmath10 of ( [ discrete mixed formulation p2 ] ) is in @xmath86 , then @xmath10 is the solution of ( [ may2 ] ) .
let @xmath87 be a shape regular partition of @xmath0 into triangles ( tetrahedra for @xmath88 ) or parallelograms ( parallelepiped for @xmath88 ) satisfying the angle condition @xcite , i.e. , there exists a constant @xmath89 such that @xmath90 where @xmath91 .
let @xmath92 be the space of polynomials of total degree at most @xmath93 if @xmath94 is a simplex , or the space of polynomials with degree at most @xmath93 for each variable if @xmath94 is a parallelogram / parallelepiped .
define the finite element spaces @xmath95 and @xmath96 by @xmath97 and @xmath98 respectively .
we introduce the mixed finite element method for problem @xmath46 : find @xmath99 such that @xmath100 for problem @xmath48 , the mixed problem reads : find @xmath101 such that @xmath102 by standard arguments , problem ( [ mixed from p1 ] ) possesses a unique solution provided there exist functions @xmath103 and @xmath104 satisfying @xmath105 then @xmath106 is the trivial solution to the system . in fact , taking @xmath107 in the first equation of ( [ mixed form p11 ] ) , one gets @xmath108 . setting @xmath109 in the second equation of ( [ mixed form p11 ] ) , one obtains @xmath110 .
similarly , it is verified that problem ( [ mixed form p2 ] ) has also a unique solution .
we define a measure of the error between the exact solution @xmath111 and the numerical solution @xmath106 by @xmath112 where @xmath113 is the standard energy norm of the numerical error @xmath114 . in this paper , we aim at robust _ a posterior _ error estimators for the numerical errors @xmath115 , @xmath116 , and @xmath117 .
we next introduce some notations that will be used later .
we denote @xmath118 the set of interior sides ( if @xmath119 ) or faces ( if @xmath88 ) in @xmath87 , @xmath120 the set of sides or faces of @xmath121 , and @xmath122 the union of all elements in @xmath87 sharing at least one point with @xmath18 . for a side or face @xmath15 in @xmath123 , which is the set of element sides or faces in @xmath87 , let @xmath124 be the diameter of @xmath15 , and @xmath125 be the union of all elements in @xmath87 sharing @xmath15 . for a function @xmath45 in the broken sobolev space
" @xmath126 , we define @xmath127|_{e } : = ( v|_{t_{+}})|_{e}-(v|_{t_{-}})|_{e}$ ] as the jump of @xmath45 across an interior side or face @xmath15 , where @xmath128 and @xmath129 are the two neighboring elements such that @xmath130 . throughout of this paper , we denote by @xmath131 a constant depending only on @xmath132 , and denote by @xmath133 constants depending on the mesh shape regularity and @xmath134 . in what follows we use the notation @xmath135 to represent @xmath136 with a generic constant @xmath22 independent of mesh size .
in addition , @xmath137 abbreviates @xmath138 .
for problem @xmath48 , @xmath29 and @xmath10 are decoupled . however , @xmath29 does not obtain any information directly from boundary conditions .
it will be difficult to develop residual - based _ a posteriori _ error estimates if the residual on the boundary is not clear . to overcome this difficulty , we shall develop a novel analytical technique ( see section 4 ) , which is based on the following approximation result .
[ julylemma1 ] let @xmath139 which satisfies @xmath140 ( operator to be understood in weak sense ) .
then it holds @xmath141 consider the functional @xmath142 minimization of such functional in @xmath78 immediately leads to the following variational problem : find @xmath143 such that @xmath144 integrating by parts , we arrive at @xmath145 which implies @xmath146 so @xmath147 is the solution to the following problem : @xmath148 since the problem ( [ july2 ] ) has only trivial solution , we have @xmath149 which completes the proof .
[ julylemma2 ] if there holds the following relation @xmath150 then @xmath45 satisfies @xmath140 . since @xmath151 , the condition @xmath152 can be satisfied when @xmath153 . on the other hand , @xmath147 is the solution to the variational problem : find @xmath143 such that @xmath154 by integrating by parts
, we obtain @xmath155 which leads to @xmath156 , this means @xmath157 [ julylemma3 ] let @xmath158 such that @xmath159 .
then there exists @xmath160 such that @xmath161 we only prove the case @xmath139 using proof by contradiction , since @xmath162 is obvious .
assume that there does not exist @xmath160 such that @xmath163 which means @xmath164 for all @xmath160 .
from the proof of lemma [ julylemma1 ] , there exists @xmath165 such that @xmath166 in particular , for any @xmath167 , let @xmath168 , it holds that @xmath169 since @xmath170 is the upper bound for @xmath171 with respect to the positive integer number @xmath172 , and @xmath173 is its supremum .
therefore , @xmath174 which means @xmath175 this leads to @xmath176 , we have @xmath177 , this leads to a contradiction .
we complete the proof .
note that we can not prove that @xmath178 is dense in @xmath78 by recursion by using lemma [ julylemma3 ] , because of @xmath179 .
denote by @xmath180 the quasi - interpolation operator of clment ( cf .
@xcite ) . [ year0 ] for all @xmath181 ,
define @xmath182 and @xmath183 the weighted factors by @xmath184 respectively .
then the following local error estimates hold for @xmath185 : @xmath186 and @xmath187 following the line of the proof of lemma 3.2 in @xcite , we obtain the desired estimates ( [ year1 ] ) and ( [ year2 ] ) . for @xmath188 , t\in\mathcal{t}_{h},e\in\mathcal{e}_{h}$ ] , denote @xmath189 and @xmath190 the two bubble functions defined in @xcite , and @xmath191 a continuation operator introduced in @xcite by @xmath192 which maps polynomials onto piecewise polynomials of the same degree .
[ year3 ] the following estimates hold for all @xmath193 ( the set of polynomials of degree at most @xmath93 ) and @xmath121 @xmath194 @xmath195 @xmath196 furthermore , for @xmath197 , set @xmath198 .
then there hold the following estimates for all @xmath197 and @xmath199 .
@xmath202 following the line of the proof of lemma 3.3 in @xcite , we attain ( [ year4])-([year9 ] ) .
for all @xmath121 , define @xmath203 and @xmath204 the elementwise indicators of @xmath29 and @xmath10 , respectively , by @xmath205 \big\|_{e}^{2 } \big\}^{1/2}\ ] ] and @xmath206 \big\|_{e}^{2 } \big\}^{1/2}.\ ] ] let @xmath207 and @xmath99 be the solutions to ( [ discrete mixed formulation p1 ] ) and ( [ mixed from p1 ] ) , respectively . then there exist positive constants @xmath208 , and @xmath209 , independent of the mesh - size function @xmath210 and @xmath11 , such that @xmath211 @xmath212 @xmath213 from the definition of the measure @xmath117 , ( [ year12 ] ) follows from ( [ year10 ] ) and ( [ year11 ] ) .
we need to prove ( [ year10 ] ) and ( [ year11 ] ) .
we have from the first equations of ( [ discrete mixed formulation p1 ] ) and ( [ mixed from p1 ] ) that @xmath214 for any @xmath215 , let @xmath216 be the clemnt interpolation of @xmath217 in @xmath96 , i.e. , @xmath218 . applying integration by parts and ( [ year16 ] ) , we get @xmath219\big\|_{e } \|\varphi-\varphi_{h}\|_{e } \big\}.\end{aligned}\ ] ] notice that @xmath220 the first estimate ( [ year10 ] ) follows from a combination of ( [ year18 ] ) , ( [ year17 ] ) , and ( [ year1])-([year2 ] ) .
we next prove ( [ year11 ] ) . from the second equation of ( [ discrete mixed formulation p1 ] ) and ( [ mixed from p1 ] ) , we get @xmath221 similarly , we have , for any @xmath162 and @xmath222 , @xmath223 \big\|_{e}\|v - v_{h}\|_{e}\big\ } \\ \label{year20 } & + \|\psi-\psi_{h}\|\|v\|.\end{aligned}\ ] ] recall the following estimates on clemnt interpolation ( cf .
@xcite ) : @xmath224 and @xmath225 for any @xmath162 , the poincar inequality implies @xmath226 a combination of ( [ year20 ] ) and ( [ year20+])-([year21 ] ) yields @xmath227\big\|_{e}^{2 } \big ) + \|\psi-\psi_{h}\| \big\ } |v|_{1}.\end{aligned}\ ] ] notice that @xmath228 ) , ( [ year21 + ] ) , and ( [ year10 ] ) yields @xmath229 this completes the proof of ( [ year11 ] ) .
let @xmath64 and @xmath230 be the solutions to ( [ discrete mixed formulation p2 ] ) and ( [ mixed form p2 ] ) , respectively .
if @xmath231 , then there exist positive constants @xmath232 , and @xmath233 , independent of the mesh - size function @xmath210 and @xmath11 , such that @xmath234 @xmath235 @xmath236 we have from @xmath231 @xmath237 for @xmath158 satisfying @xmath159 , from the proofs of lemmas [ julylemma1 ] and [ julylemma3 ] , there exist @xmath165 and @xmath160 , such that @xmath238 let @xmath239 be the clemnt interpolation of @xmath147 in @xmath96 . from the first equation of ( [ discrete mixed formulation p2 ] ) and ( [ mixed form p2 ] )
, we have @xmath240 from ( [ year25 ] ) , we have @xmath241 repeating the proof of ( [ year17 ] ) , and applying ( [ year1])-([year2 ] ) , we have @xmath242\big\|_{e}\|\tilde{v}-\tilde{v}_{h}\|_{e}\big\}\vspace{2mm}\\ & \ & \leq\displaystyle c\big\{\sum\limits_{t\in\mathcal{t}_{h}}\eta_{\psi , t}^{2}\big\}^{1/2}\|\tilde{v}\|_{\mathcal{e}}. \end{array}\ ] ] using the triangle inequality and ( [ year24 ] ) , we have @xmath243 a combination of ( [ year25 + ] ) , ( [ year24 ] ) , ( [ year26 ] ) , and ( [ year27 ] ) yields @xmath244 from ( [ year23 ] ) and ( [ year28 ] ) , we obtain @xmath245 which leads to the desired estimate ( [ year13 ] ) . repeating the proof of ( [ year11 ] ) and ( [ year12 ] ) , we obtain ( [ year14 ] ) and ( [ year15 ] ) . the condition @xmath231 is usually satisfied , since @xmath246 is the residual , which does nt vanish in usual . here
@xmath247 is the piecewise laplacian of @xmath103 .
in this section , we analyze the efficiency of the _ a posteriori _ error estimates developed in section 4 . to avoid the appearance of high order term
, we assume that @xmath248 is a piecewise polynomial .
[ year29 ] for all @xmath121 , there hold @xmath249 and @xmath250 we first prove ( [ year30 ] ) . to this end , let @xmath251 .
recall the bubble function @xmath189 introduced in section 3 . from ( [ year4 ] ) , integration by parts , and ( [ year6 ] )
, we have @xmath252 the desired estimate ( [ year30 ] ) follows .
we next prove ( [ year31 ] ) . for convenience , denote @xmath253 .
similarly , we have from @xmath27 that @xmath254 applying inverse estimate and ( [ year5 ] ) , we have @xmath255 the estimate ( [ year31 ] ) follows immediately . [ year32 ] for all @xmath256 , there hold @xmath257\big\|_{e } \lesssim\|\psi-\psi_{h}\|_{\mathcal{e},\omega_{e}}\ ] ] and @xmath258 \big\|_{e } \lesssim|u - u_{h}|_{1,\omega_{e } } + h_{e}\|\psi-\psi_{h}\|_{\omega_{e}}.\ ] ] we first prove ( [ year33 ] ) . to this end , let @xmath259 $ ] .
recall the bubble function @xmath260 and the extension operator @xmath191 introduced in section 3 .
let @xmath261 .
an application of integration by parts leads to @xmath262,v_{e } \big)_{e } , \end{array}\ ] ] where @xmath263 is the elementwise laplace operator .
a combination of the above equality and ( [ year7])-([year9 ] ) leads to @xmath264,v_{e } \big)_{e}\vspace{2mm}\\ & = & ( f+\varepsilon^{2}\triangle_{h}\psi_{h}-\psi_{h},v_{e})_{\omega_{e}}\vspace{2mm}\\ & \ & \ \ -(\varepsilon^{2}\nabla(\psi-\psi_{h}),\nabla v_{e})_{\omega_{e}}-(\psi-\psi_{h},v_{e})_{\omega_{e}}\vspace{2mm}\\ & \lesssim&\|f+\varepsilon^{2}\triangle_{h}\psi_{h}-\psi_{h}\|_{\omega_{e}}\|v_{e}\|_{\omega_{e}}+ \|\psi-\psi_{h}\|_{\mathcal{e},\omega_{e}}\|v_{e}\|_{\mathcal{e},\omega_{e}}\vspace{2mm}\\ & \lesssim&\varepsilon^{1/2}\min\{1,h_{e}\varepsilon^{-1}\}^{1/2}\|f+\varepsilon^{2}\triangle_{h}\psi_{h}- \psi_{h}\|_{\omega_{e}}\|\sigma\|_{e}\vspace{2mm}\\ & \ & \
\ + \|\psi-\psi_{h}\|_{\mathcal{e},\omega_{e}}\varepsilon^{1/2}\min\{1,h_{e}\varepsilon^{-1}\}^{-1/2}\|\sigma\|_{e}. \end{array}\ ] ] by the definition of @xmath183 for @xmath265 and the local shape regularity of the mesh , we obtain from the above inequality that @xmath266 in the last step , estimate ( [ year30 ] ) is used .
we complete the proof of ( [ year33 ] ) .
we next prove ( [ year34 ] ) . for convenience , denote @xmath267 $ ] and @xmath268 , where @xmath269 for @xmath270 .
similarly , we have @xmath271,v_{e } \big)_{e},\ ] ] which leads to the following estimate : @xmath272,v_{e } \big)\vspace{2mm}\\ & = & ( \psi+\triangle_{h}u_{h},v_{e})_{\omega_{e}}-(\nabla(u - u_{h}),\nabla v_{e})_{\omega_{e}}\vspace{2mm}\\ & = & ( \psi-\psi_{h},v_{e})_{\omega_{e}}+(\triangle_{h}u_{h}+\psi_{h},v_{e})_{\omega_{e}}-(\nabla(u - u_{h}),\nabla v_{e})_{\omega_{e}}\vspace{2mm}\\ & \lesssim&\|\psi-\psi_{h}\|_{\omega_{e}}h_{e}^{1/2}\|\sigma\|_{e}+h_{e}^{1/2}\|\triangle_{h}u_{h } + \psi_{h}\|_{\omega_{e}}\|\sigma\|_{e}\vspace{2mm}\\ & \ & \ + |u - u_{h}|_{1,\omega_{e}}h_{e}^{-1/2}\|\sigma\|_{e}. \end{array}\ ] ] we obtain from the above inequality that @xmath273 in the last step above , we employ the estimate ( [ year31 ] ) .
we complete the proof of ( [ year34 ] ) .
let @xmath207 and @xmath99 be the solutions to ( [ discrete mixed formulation p1 ] ) and ( [ mixed from p1 ] ) , respectively .
then there exist positive constants @xmath274 and @xmath275 , independent of the mesh - size function @xmath210 and @xmath11 , such that @xmath276 and @xmath277 summing ( [ year30 ] ) and ( [ year33 ] ) over all @xmath121 , we obtain the first estimate .
similarly , we get the second one
. let @xmath64 and @xmath230 be the solutions to ( [ discrete mixed formulation p2 ] ) and ( [ mixed form p2 ] ) , respectively .
then there exist positive constants @xmath278 and @xmath279 , independent of the mesh - size function @xmath210 and @xmath11 , such that @xmath280 and @xmath281 these two estimates follow from lemmas [ year29 ] and [ year32 ] .
theorems 4.1 and 5.3 ( and theorem 4.2 and 5.4 ) indicate that the ratios between the upper and lower bounds , i.e. , @xmath282 and @xmath283 ( and @xmath284 and @xmath285 ) , do not depend on the singular perturbation parameter @xmath11 .
therefore , the estimators developed in this paper are fully robust with respect to @xmath11 .
this further implies that each component of the new measure of the error is balanced with respect to the perturbation parameter .
in this section , we test our _ a posteriori _ error estimators on two model problems .
note that all programs were developed by ourselves . consider problem ( [ pde1 ] ) and ( [ boundary condition2 ] ) on the unit square @xmath286 . we suppose the exact solution of this model has the form @xmath287 the function @xmath10 has a boundary layer , which varies significantly near @xmath288 . .
lower : the mesh after 10 iterations with 478 triangles ( left ) and the mesh after 12 iterations with 1010 triangles ( right ) , generated by the elementwise indicator @xmath289 . here
@xmath290 and @xmath291 for all plots.,width=432 ] our initial mesh consists of eight isosceles right triangles .
we employ drfler marking strategy @xcite with the marking parameter @xmath291 and use the longest edge " refinement @xcite to obtain an admissible mesh .
plots in figure [ fig1 ] depict the estimators of @xmath292 ( upper and middle ) , and @xmath293 ( lower ) , respectively .
we observe that strong mesh refinement near the line @xmath288 , which indicates the estimators of the errors @xmath115 and @xmath117 capture boundary layers well .
figure [ fig4 ] demonstrates finite element approximations to @xmath10 ( left ) and @xmath27 ( right ) .
it is observed that the function @xmath10 does nt possess layer , and that @xmath29 has boundary layer near @xmath288 .
on the other hand , the upper two plots of figure [ fig6 ] display the estimated and exact errors for @xmath115 ( left ) and @xmath117 ( right ) , respectively .
it is observed that the estimated convergence curve overlaps the curve of @xmath115 , which indicates that the estimator for @xmath115 is asymptotically exact even for very small @xmath11 .
we also observe that the estimated convergence curve is parallel to the curve @xmath117 independent of @xmath290 , and both curves decrease in optimal rates .
note that the study of convergence and optimality of adaptive algorithms is still in its infancy , and has been carried out mainly for standard adaptive finite element method for general second order elliptic problems ; see , e.g. , @xcite .
( left ) and @xmath117 ( right ) against the number of elements in adaptively refined meshes for @xmath290 .
lower : exact errors of @xmath115 ( left ) and @xmath117 ( right ) against the number of elements in adaptively refined meshes for @xmath294 and @xmath290 . here
the marking parameter @xmath291.,width=480 ] the two lower plots of figure [ fig6 ] depict error curves for @xmath115 ( left ) and @xmath117 ( right ) , respectively .
it is observed that the convergence curves for @xmath294 and @xmath290 are consistent , which indicates that the errors reduce uniformly with respect to @xmath11 .
in addition , we include in figure [ fig6 ] an optimal theoretical convergence line with slope @xmath295 .
the plots indicate that @xmath115 and @xmath117 decrease in the optimal convergence rates .
tables [ tab : egone1 ] and [ tab : egone2 ] show some results of the actual errors @xmath115 and @xmath117 , the _ a posteriori _ indicators @xmath296 and @xmath297 , and the effectivity indexes eff - index@xmath298 for @xmath29 and eff - index@xmath299 for @xmath111 for example 1 , where eff - index@xmath300 , eff - index@xmath301 .
it is observed that the effectivity indices of the error @xmath115 are close to 1 , and that the effectivity indices of the error @xmath115 are about 1.5 .
this suggests that our estimators are robust with respect to @xmath11 .
.example 1 : @xmath93 number of iterations ; @xmath296 numerical result of estimated error for @xmath115 ; eff - index@xmath298 the corresponding effectivity index for @xmath29 ( the ratio of estimated and exact errors ) .
here @xmath294 , @xmath302 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] table [ egtwo1 ] reports the given tolerance tol , the number of iterations @xmath93 , the estimated error ( @xmath303 ) for @xmath115 , the degrees of freedom dof , the smallest mesh size @xmath304 for example 2 , which show that the required dof depends on both tol and @xmath11 , and that the layer is gradually resolved , because the smallest mesh size @xmath305 has arrived at the magnitude of @xmath11 after 22 iterations .
figure [ fig11 ] shows the estimated errors of @xmath115 ( left ) and @xmath117 ( or @xmath116 ) ( right ) , respectively .
we observe again that the estimated errors reduce uniformly with respect to @xmath11 in both norms with almost optimal rate @xmath295 .
g. awanou , _ robustness of a spline element method with constraints .
36 ( 2008 ) , pp .
421 - 432 .
i. babuka , j. osborn , and j. pitkranta , _ analysis of mixed methods using mesh dependent norms .
comput . , 35 ( 1982 ) , pp .
1039 - 1062 .
bank and a. weiser , _ some a posteriori error estimators for elliptic partial differential equations . _ math .
44 ( 1985 ) , pp .
283 - 301 .
a. charbonneau , k. dossou , and r. pierre , _ a residual - based a posteriori error estimator for the ciarlet - raviart formulation of the first biharmonic problem .
methods partial differential equations , 13 ( 1997 ) , pp . 93 - 111
. l. chen , m. holst , and j. xu , _ convergence and optimality of adaptive mixed finite element methods .
comput . , 78 ( 2009 ) , pp .
z. chen and r.h .
nochetto , _ residual type a posteriori error estimates for elliptic obstacle problems .
84 ( 2000 ) , pp .
527 - 548 .
p. danumjaya and a.k .
pani , _ mixed finite element methods for a fourth order reaction diffusion equation .
meth . part .
d. e. , 28 ( 2012 ) , pp .
1227 - 1251 .
a. demlow and n. kopteva , _ maximum - norm a posteriori error estimates for singularly perturbed elliptic reaction - diffusion problems .
_ numer . math .
133 ( 2016 ) , pp .
707 - 742 .
falk and j.e .
osborn , _ error estimates for mixed methods . _ m2an math . model .
, 14 ( 1980 ) , pp .
249 - 277 .
l. s. fank , _ singular perturbations in elasticity theory , analysis and its applications .
_ 1 . , ios press , amsterdam , 1997 .
t. gudi , _ residual - based a posteriori error estimator for the mixed finite element approximation of biharmonic equation . _ numer .
methods partial differential equations , 27 ( 2011 ) , pp .
315 - 328 .
j. guzmn , d. leykekhman , and m. neilan , _ a family of non - conforming elements and the analysis of nitsche s method for a singularly perturbed fourth order problem .
_ calcolo , 49 ( 2012 ) , pp .
h. han and z. huang , _ an equation decomposition method for the numerical solution of a fourth - order elliptic singular perturbation problem .
methods partial differential equations , 28 ( 2012 ) , pp .
942 - 953 .
j. hu and z.c .
shi , _ a new a posteriori error estimate for the morley element .
, 112 ( 2009 ) , pp .
r. lin and m. stynes , _ a balanced finite element method for singularly perturbed reaction - diffusion problems .
_ siam j. numer .
, 50 ( 2012 ) , pp .
2729 - 2743 .
p. morin , r.h .
nochetto , and k.g .
siebert , _ convergence of adaptive finite element methods .
_ siam review , 44 ( 2002 ) , pp .
631 - 658 .
morley , _ the triangular equilibrium element in the solutionof plate bending problems . _
aero . quart .
, 19 ( 1968 ) , pp .
149 - 169 .
b. semper , _ conforming finite element approximations for a fourth - order singular perturbation problem .
_ siam j. numer .
, 29 ( 1992 ) , pp .
1043 - 1058 .
z. shi , _ error estimates of morley element .
sinica , 12 ( 1990 ) , pp .
113 - 118 .
m. wang and s. zhang , _ local a priori and a posteriori error estimates of finite elements for biharmonic equation .
_ research report , 13 ( 2006 ) , school of mathematical science and institute of mathematics , peking university .
s. zhang and z.m .
zhang , _ invalidity of decoupling a biharmonic equation to two piosson equations on non - convex polygons .
_ international journal of numerical analysis and modeling , 5 ( 2007 ) , pp . | we consider mixed finite element approximation of a singularly perturbed fourth - order elliptic problem with two different boundary conditions , and present a new measure of the error , whose components are balanced with respect to the perturbation parameter .
robust residual - based _ a posteriori _ estimators for the new measure are obtained , which are achieved via a novel analytical technique based on an approximation result .
numerical examples are presented to validate our theory
. key words .
fourth order elliptic singularly perturbed problems , mixed finite element methods , a new measure of the error , robust residual - based _ a posteriori _ error estimators ams subject classifications . 65n15 , 65n30 , 65j15 |
although quantum fluctuations of the inflaton are often taken to be responsible for the origin of density perturbations , other mechanisms such as the curvaton @xcite , where fluctuations of a scalar field other than the inflaton generate primordial perturbations , have also attracted much attention recently . in particular , in the light of the recent result from wmap5 which suggests that primordial non - gaussianity may be large @xcite .
the current constraint on @xmath3 is @xmath12 in ref .
@xcite or @xmath13 at @xmath14 c.l . in ref .
although purely gaussian fluctuations with @xmath15 are allowed , the central value is away from zero . ]
, the curvaton mechanism may be attractive since a large primordial non - gaussianity can be generated in this scenario @xcite , whereas simplest inflation models predict a non - linearity parameter @xmath3 that is of the order of the slow - roll parameters ( or o(1 ) at most ) , and hence practically imply a gaussian perturbation .
usually the curvaton potential is assumed to be quadratic .
however , there is no other reason for this except simplicity , and in fact in any realistic particle physics model the curvaton can be expected to have some self - interactions .
thus , deviations from the exact quadratic form are also worth investigating .
they have been discussed in @xcite , where it has been pointed out that non - quadratic contributions to the potential can modify the resultant curvaton perturbations in a significant manner . in particular , the prediction for the non - linearity parameters @xmath3 and @xmath4 can change considerably as compared to the quadratic case .
in addition , there is yet another assumption which is tacitly adopted in the curvaton literature : the background evolution of the universe is determined by radiation . with this assumption ,
the curvaton starts to oscillate during a radiation - dominated ( rd ) epoch .
if the curvaton decays before dominating the universe , radiation is always the dominant component and controls the background evolution of the universe .
however , it is also possible that after inflation , the inflaton is oscillating around the minimum of the potential for a while and that the curvaton begins to oscillate during such epoch . in this case , the background evolution is different from the case of radiation and is determined by a matter - like component if the inflaton potential is approximatively quadratic .
after inflation there could also exist a possibility of a kination - dominated phase where the kinetic term of some scalar field can dominate the energy density of the universe .
such fluid has a stiff equation of state with @xmath16 while its energy density decreases as @xmath17 . in this paper , we investigate how the background evolution of the universe affects the curvature perturbation generated from the curvaton .
we first show that when the curvaton potential has a quadratic form , the background evolution has little effect on the curvature perturbation .
however , when the curvaton potential includes a non - quadratic term , the nonlinear evolution of the curvaton field is affected much by the background , and the resultant curvature perturbations can be significantly modified from the usual rd case .
we investigate this issue by assuming a general background fluid and a potential that slightly deviates from a quadratic form , and derive the dependence of the curvature perturbation and/or the non - linearity parameters such as @xmath3 and @xmath4 on different background fluids .
the structure of the paper is as follows . in the next section ,
we summarize the formalism and give definitions of the quantities required for the subsequent analysis . in section [ sec : effect ] , we discuss how the background evolution affects the quantities such as power spectrum and non - linearity parameters , using the appropriate formulas for arbitrary background fluids .
the final section is devoted to a summary and a discussion of the results .
let us begin by summarizing the formalism and the definitions of the various quantities required for the subsequent discussion . here
we consider a potential of the curvaton field @xmath18 which , in addition to the usual quadratic term , also includes a non - renormalizable term : @xmath19 where @xmath20 is the mass of the curvaton , and @xmath21 is a constant . for the purpose of this paper , it is enough to investigate the case of a slight deviation from the purely quadratic form .
thus in the following we assume that the quadratic term always dominates over the non - quadratic term ( for a general discussion of the ramifications of non - quadratic terms in curvaton models , see @xcite ) .
we characterize the relative contribution of the non - quadratic term at the time when the curvaton is still in a slowly - rolling regime by the parameter @xmath22 , defined as @xmath23 to investigate the curvature perturbation @xmath24 generated by the curvaton field , we adopt the @xmath25 formalism @xcite and calculate @xmath24 up to the third order as @xmath26 once we obtain @xmath24 up to the third order , the power spectrum @xmath27 , bispectum @xmath28 , and trispectrum @xmath29 are given by @xmath30 @xmath31 @xmath32 where @xmath33 and @xmath34 can be written as @xmath35 here @xmath36 and @xmath4 are non - linearity parameters often used in the literature .
note that , in our case , @xmath37 is related to @xmath3 by @xmath38 thus , alternatively , we may assume the following expansion as the definition of @xmath3 and @xmath4 @xmath39 where @xmath40 denotes the curvature perturbation at linear order .
let us now derive the formulae for the power spectrum , @xmath3 and @xmath4 for a general background equation of state . as we mentioned in the introduction , usually one assumes that the curvaton oscillates in a radiation background evolution .
however , it is obvious that the inflaton can be expected to oscillate around the minimum of its ( quadratic ) potential for some time before it decays . for a weakly coupled inflaton field
, the duration of this epoch could be fairly long . during that time the inflaton behaves like matter ,
and it is possible that the curvaton oscillations begin during this epoch .
eventually the inflaton will decay into radiation ; this can well happen before the curvaton oscillations dominate the energy budget of the universe , or before the curvaton decay , whichever happens first .
hence the universe becomes radiation - dominated so that the background evolution for oscillating curvaton is controlled first by a matter component , followed by radiation domination .
having this kind of situation in mind , let us assume there is a transition in the background from one fluid to another one at time @xmath41 .
we then parametrize the energy density of the background fluid before the transition time by @xmath42 whereas after the transition , for @xmath43 , we have @xmath44 for the case of oscillating inflaton background , the transition epoch @xmath41 corresponds to the time when the inflaton decays . in this case , for @xmath45 , @xmath46 while for @xmath47 , @xmath48 .
another example is a universe that is first kination - dominated so that the kinetic energy of a scalar field dominates the energy density of the universe .
then , after some time , the universe becomes radiation - dominated . in this case ,
@xmath49 for @xmath45 while after the transition time @xmath50 . with this parametrization
, we obtain the curvature perturbation at the linear order as @xmath51 where @xmath52 the prime denotes the derivative with respect to @xmath53 . here
@xmath54 and @xmath55 roughly correspond to the fraction of energy density of the curvaton to the total energy density at the time of the curvaton decay and the transition of the background , respectively .
their precise definitions are @xmath56 furthermore , @xmath57 is defined by @xmath58 by calculating the curvature perturbation up to the third order , we find that the non - linearity parameters @xmath3 and @xmath4 read as @xmath59,\ ] ] @xmath60,\ ] ] here the derivatives of @xmath61 with respect to @xmath62 appear .
they are given in terms of the derivatives of @xmath54 and @xmath55 with respect to @xmath62 as @xmath63 and @xmath64 the derivatives @xmath65 and @xmath66 are explicitly given by @xmath67 ( 1 -k r_{\rm tr } ) , \\
\frac{d r_{\rm tr}}{d \sigma_{\rm osc } } & = & \frac{2}{3\sigma_{\rm osc } } r_{\rm tr } ( 1-r_{\rm tr } ) [ 3 + ( \alpha-3 ) r_{\rm tr } ] \end{aligned}\ ] ] the second derivatives of @xmath54 and @xmath55 with respect to @xmath62 can be derived by differentiating the above equations and they are calculated as @xmath68 - \frac{2k}{3\sigma_{\rm osc } } \frac{d r_{\rm tr}}{d \sigma_{\rm osc } } r_{\rm dec } ( 1-r_{\rm dec } ) \ { 3 + ( \beta-3 ) r_{\rm dec } \ } , \notag \\
\\ \notag \\
\frac{d^2 r_{\rm tr}}{d \sigma_{\rm osc}^2 } & = & \frac{2}{3\sigma_{\rm osc } } \frac{d r_{\rm tr}}{d \sigma_{\rm osc } } \left [ -\frac{3}{2 } + ( 1 - 2 r_{\rm tr } ) \ { 3 + ( \alpha-3 ) r_{\rm tr } \ } + ( \alpha-3 ) r_{\rm tr } ( 1 - r_{\rm tr } ) \right].\end{aligned}\ ] ] we can now write down @xmath3 and @xmath4 as functions of the parameters @xmath69 and @xmath55 in an explicit way , although in general the expressions are very complicated . in fact , as far as we consider only the cases with @xmath70 , @xmath55 should always be smaller than @xmath54 and in most cases we may assume that @xmath71 . in this case ,
@xmath72 and @xmath73 and @xmath4 may approximately be written as @xmath74 .
\notag \\\end{aligned}\ ] ] as one can notice from these expressions , when @xmath54 is small , which is the interesting case because then non - gaussianity can be large , the effects of the background are almost encoded in the changes of @xmath62 and its derivatives .
the quantities such as @xmath75 and @xmath76 are evaluated at the beginning of the curvaton oscillation , and thus the background evolution after the transition is irrelevant if we assume a nearly quadratic potential for the curvaton .
thus in this case , after its change the background evolution does not affect much the curvature perturbation but modifies @xmath40 and its non - linearity parameters at most of order @xmath77 .
this can be seen directly from the above expressions : @xmath78 affects the coefficients by @xmath77 .
if there is no change in the background evolution up to or until after the time at which the curvaton decays , the expressions for the curvature perturbation and non - linearity parameters are quite similar to the case with @xmath79 mentioned above . in this case , we can simply set @xmath80 and @xmath81 in the above equations
. then @xmath82 and @xmath4 can be found by replacing @xmath78 with @xmath2 in eqs . , and .
let us now apply the formalism of the previous section and discuss how the background evolution affects the amplitude of the curvature perturbation and its non - linearity parameters .
as we pointed out , the transition in the background does not affect the perturbation much if the transition occurs much before curvaton decay .
hence here we focus on the case of no transition in the background , i.e. , a single fluid controls the background evolution from the time well before the curvaton begins to oscillate to well after the curvaton has decayed . when there is no transition in the background evolution we can adopt the formulae eqs . ,
and by replacing @xmath78 by @xmath2 in the equations , as mentioned above . to appreciate the impact of background evolution on the curvature perturbation and its nonlinearity , we plot @xmath82 and @xmath4 in figs .
[ fig : fnl_gnl_n][fig : fnl_gnl ] . in fig .
[ fig : fnl_gnl_n ] , the values of @xmath3 and @xmath4 are shown as a function of @xmath83 ( see eq .
( [ eq : def_s ] ) for the definition of the dimension of the non - quadratic contribution to the potential ) for the cases of @xmath84 and @xmath85 which correspond to matter , radiation and kination background , respectively , in our notation .
as seen from the figures , the larger the non - quadratic power @xmath83 is , the more pronounced the effect of the background becomes . to observe the magnitude of the effect more quantitatively , the values of @xmath82 and @xmath4 relative to those for the case with @xmath86 and @xmath87
are plotted in figs .
[ fig : zeta_1][fig : gnl_1 ] for several values of @xmath83 as a function of @xmath2 . in these figures
, we fix the relative strength of the non - quadratic part @xmath22 and @xmath54 to @xmath88 and @xmath89 . in figs .
[ fig : zeta_2][fig : gnl_2 ] we also plot the same information but this time for several values of @xmath22 , fixing @xmath90 and @xmath54 .
interestingly , when the potential of the curvaton is close to a purely quadratic one , the background evolution does not affect the results much
. however , whenever the potential deviates from the quadratic form , the background tends to suppress the curvature perturbation @xmath40 the more non - quadratic the potential is .
it should also be noted that when @xmath2 increases , or the background fluid becomes more stiff , @xmath40 becomes suppressed . for
fixed @xmath22 and @xmath83 , the stiffening of the background fluid drives @xmath3 and @xmath91 to increasingly negative territory .
a point worth stressing is that these modifications are not small but e.g. by changing of the radiation dominated background to matter dominated one induces shifts in the non - linearity parameters that in principle could be easily observable .
finally , let us comment on the relation between @xmath3 and @xmath4 .
as pointed out in @xcite , when the potential of the curvaton has a purely quadratic form and @xmath92 is small , @xmath3 and @xmath4 are related as @xmath93 however , when the potential deviates from a quadratic form , the consistency relation " between @xmath3 and @xmath4 becomes @xmath94 where @xmath95 is assumed . because of the nonlinear evolution of @xmath18 , which is encoded in @xmath62 and its derivatives appearing on the right hand side of eq . , the nonlinearity parameters are now related as @xmath96 .
this is in sharp contrast to the quadratic case where @xmath97 .
furthermore , since the coefficient of the relation depends on the nonlinear evolution of @xmath18 , it can be affected by the form of the potential and the background evolution . in this respect ,
the comparison of @xmath3 and @xmath98 , if they ever are observed has been obtained for the case of a negligible @xmath3 ; the limit is @xmath99 @xcite . ] , can be very useful for probing the form of the curvaton potential and the equation of state of the background during the time when the curvaton fluctuations generate the curvature perturbation . to demonstrate this explicitly , in fig .
[ fig : fnl_gnl ] we plot the value of @xmath4 as a function of @xmath3 for several values of @xmath2 with @xmath90 and @xmath100 . for reference , we also plot the case of the pure quadratic potential with rd background @xmath101 .
when @xmath3 is large , which corresponds to the case of @xmath95 , the above relations hold . as a consequence , as one can see in fig .
[ fig : fnl_gnl ] , there is a definite difference in the @xmath102 relation for different backgrounds even when the non - quadratic term in the curvaton potential remains the same .
thus the trispectrum may also provide important information not only about the curvaton potential but also about the equation of state of the background fluid during curvaton oscillations .
in this paper , we have investigated the effect of the background evolution on the curvaton non - gaussianity , assuming a curvaton potential which slightly deviates from the quadratic one .
such a study is motivated by the possibility that after inflation , the inflaton keeps oscillating about its global minimum for a long time so that the curvaton could actually decay while the universe is still effectively matter dominated .
more exotic temporary possibilities , such as a kination driven universe , could also be envisaged .
therefore we have considered an ideal background fluid with some generic equation of state , leading to a background evolution @xmath103 , with @xmath2 a free parameter .
it turns out that the changing of the background to radiation , or equivalently , a change in the value of @xmath2 , that takes place during curvaton oscillations , has by itself little effect on the perturbation .
what matters is the nature of the background evolution before radiation domination finally kicks in , and as we show , it can lead to significant and potentially observable consequences .
the non - linearity parameters @xmath3 and @xmath4 , as well as the linear curvature perturbation @xmath40 , depend on the nature of the background fluid .
we find that the dependence on the background fluid becomes more pronounced as the deviation of the curvaton potential from the quadratic one increases .
typically , when replacing one background fluid with another , one induces effects on @xmath3 that are easily of the order of @xmath104 but could also be much larger , depending on the relative strength of the non - quadratic part of the potential .
an interesting issue is the relation between @xmath3 and @xmath91 .
we have showed that the relation depends both on the form of the curvaton potential and the background evolution , which we find a rather surprising result . hence measuring
both @xmath3 and @xmath4 , or both the bispectrum and the trispectrum , would yield information not only on the curvaton self - interactions but also on the equation of state of the background fluid . since that is linked to dynamics in the inflaton sector , here
arises a possibility of probing inflaton physics at the very end of inflation .
* acknowledgments : * t.t . would like to thank the helsinki institute of physics for the hospitality during the visit , where this work was initiated .
this work is supported in part the grant - in - aid for scientific research from the ministry of education , science , sports , and culture of japan no.19740145 ( t.t . ) , and in part by the academy of finland grant 114419 ( k.e . ) .
k. a. malik and d. h. lyth , jcap * 0609 * , 008 ( 2006 ) [ arxiv : astro - ph/0604387 ] . m. sasaki , j. valiviita and d. wands , phys .
d * 74 * , 103003 ( 2006 ) [ arxiv : astro - ph/0607627 ]
. q. g. huang , arxiv:0801.0467 [ hep - th ] .
k. ichikawa , t. suyama , t. takahashi and m. yamaguchi , phys .
d * 78 * , 023513 ( 2008 ) [ arxiv:0802.4138 [ astro - ph ] ] . t. multamaki , j. sainio and i. vilja , phys .
d * 79 * , 103516 ( 2009 ) [ arxiv:0803.2637 [ astro - ph ] ] .
k. enqvist and t. takahashi , jcap * 0809 * , 012 ( 2008 ) [ arxiv:0807.3069 [ astro - ph ] ] .
q. g. huang and y. wang , jcap * 0809 * , 025 ( 2008 ) [ arxiv:0808.1168 [ hep - th ] ]
. q. g. huang , jcap * 0811 * , 005 ( 2008 ) [ arxiv:0808.1793 [ hep - th ] ] .
p. chingangbam and q. g. huang , jcap * 0904 * , 031 ( 2009 ) [ arxiv:0902.2619 [ astro-ph.co ] ] . k. dimopoulos , g. lazarides , d. lyth and r. ruiz de austri , phys .
d * 68 * ( 2003 ) 123515 [ arxiv : hep - ph/0308015 ] .
k. enqvist , s. nurmi , g. rigopoulos , o. taanila and t. takahashi , arxiv:0906.3126 [ astro-ph.co ] .
g. dvali , a. gruzinov and m. zaldarriaga , phys .
d * 69 * , 023505 ( 2004 ) [ arxiv : astro - ph/0303591 ] .
l. kofman , arxiv : astro - ph/0303614 .
m. zaldarriaga , phys .
d * 69 * , 043508 ( 2004 ) [ arxiv : astro - ph/0306006 ] .
t. suyama and m. yamaguchi , phys .
d * 77 * , 023505 ( 2008 ) [ arxiv:0709.2545 [ astro - ph ] ] .
k. ichikawa , t. suyama , t. takahashi and m. yamaguchi , phys .
d * 78 * , 063545 ( 2008 ) [ arxiv:0807.3988 [ astro - ph ] ]
. t. takahashi , m. yamaguchi and s. yokoyama , arxiv:0907.3052 [ astro-ph.co ] .
a. a. starobinsky , jetp lett .
* 42 * ( 1985 ) 152 [ pisma zh .
eksp . teor .
* 42 * ( 1985 ) 124 ] .
m. sasaki and e. d. stewart , prog .
phys . *
95 * , 71 ( 1996 ) [ arxiv : astro - ph/9507001 ] .
m. sasaki and t. tanaka , prog .
* 99 * , 763 ( 1998 ) [ arxiv : gr - qc/9801017 ] .
d. h. lyth , k. a. malik and m. sasaki , jcap * 0505 * , 004 ( 2005 ) [ arxiv : astro - ph/0411220 ] .
v. desjacques and u. seljak , arxiv:0907.2257 [ astro-ph.co ] . | we investigate how the background evolution affects the curvature perturbations generated by the curvaton , assuming a curvaton potential that may deviate slightly from the quadratic one , and parameterizing the background fluid density as @xmath0 , where @xmath1 is the scale factor , and @xmath2 depends on the background fluid .
it turns out that the more there is deviation from the quadratic case , the more pronounced is the dependence of the curvature perturbation on @xmath2 .
we also show that the background can have a significant effect on the nonlinearity parameters @xmath3 and @xmath4 . as an example , if at the onset of the curvaton oscillation there is a dimension 6 contribution to the potential at 5 % level and the energy fraction of the curvaton to the total one at the time of its decay is at 1% , we find variations @xmath5 and @xmath6 between matter and radiation dominated backgrounds .
moreover , we demonstrate that there is a relation between @xmath3 and @xmath4 that can be used to probe the form of the curvaton potential and the equation of state of the background fluid . hip-2009 - 23/th + .5 in * effect of background evolution + on the curvaton non - gaussianity * .45 in kari enqvist@xmath7 and tomo takahashi@xmath8 .45 in _
@xmath9 helsinki institute of physics , university of helsinki , po box 64 , fin-00014 , finland + @xmath10 department of physical science , university of helsinki , + po box 64 , fin-00014 , finland + @xmath11 department of physics , saga university , saga 840 - 8502 , japan _
.4 in |
since the first discovery of a transiting planet around hd 209458 @xcite , 235 transiting extrasolar systems have already been confirmed up to december 4th , 2012 . whilst transiting exoplanets offer unique scientific possibilities , their study involves several complications . in general , it is impossible to measure the mass and radius of a planet based on a dataset obtained with one observational technique .
transit light curves allow us to determine just the relative size of a star and planet , the orbital inclination and the stellar limb - darkening coefficients . by combining this with radial - velocity measurements ,
the observations offer the opportunity to measure the precise stellar and planetary parameters . in order to obtain such parameters ,
some constraints are needed , and are usually provided by forcing the properties of the host stars to match theoretical expectations @xcite .
significant uncertainties remain in the stellar mass and radius determinations of many systems . in some cases , this is due to poorly determined photospheric properties ( i.e. effective temperature and metallicity ) , and in other cases due to a lack of an accurate luminosity estimate @xcite .
in addition , the different methods used for these determinations as well as different approaches toward systematic errors are leading to rather inhomogeneous set of planet properties .
because of such inhomogenities , recently a few papers were published where authors re - analyzed a large subset of known transiting planets , and applied a uniform methodology to all systems ( e.g. @xcite ) . in this paper we focus on the transiting system tres-3 .
the system consists of a nearby g - type dwarf and a massive hot jupiter with an orbital period of 1.3 days .
it was discovered by @xcite and also detected by the superwasp survey @xcite .
later , @xcite presented new spectroscopic and photometric observations of the host star . a detailed abundance analysis based on high - resolution spectra yields [ fe / h ] = @xmath20.19 @xmath3 0.08 , @xmath4 = 5650 @xmath3 75 k and log @xmath5 = 4.0 @xmath3 0.1 .
the spectroscopic orbital solution was improved with new radial velocity measurements obtained by @xcite . moreover , these authors redetermined the stellar parameters ( i.e. @xmath6 = 0.928@xmath7 @xmath8 and @xmath9 = 0.829@xmath10 @xmath11 ) and finally , the new values of the planetary mass and radius were determined ( see tab [ tab03 ] ) .
they also studied the transit timing variations ( ttvs ) of tres-3 and noted significant outliers from a constant period . in the same year
, @xcite presented the follow - up transit photometry .
it consisted of nine transits of tres-3 , taken as part of transit timing program using the rise instrument on the liverpool telescope .
these transits , together with eight transit times published before @xcite , were used to place upper mass limit as a function of the period ratio of a potential perturbing planet and transiting planet .
it was shown that timing residuals are sufficiently sensitive to probe sub - earth mass planet in both interior and exterior 2:1 resonances , assuming that the additional planet is in an initially circular orbit . @xcite
has observed tres-3 as a part of the nasa _ epoxi _ mission of opportunity .
they detected a long - term variability in the tres-3 light curve , which may be due to star spots .
they also confirmed that the planetary atmosphere does not have a temperature inversion .
later , @xcite observed nine primary transits of the hot jupiter tres-3b in several optical and near - uv photometric bands from june 2009 to april 2012 in an attempt to detect its magnetic field .
authors determined an upper limit of tres-3b s magnetic field strength between 0.013 and 1.3 g using a timing difference of 138 s derived from the nyquist
shannon sampling theorem .
they also presented a refinement of the physical parameters of tres-3b , an updated ephemeris and its first published near - uv light curve .
the near - uv planetary radius of @xmath12 = 1.386@xmath13 @xmath14 was also determined .
this value is consistent with the planet s optical radius .
recently , @xcite observed eleven transits of tres-3b over a two year period in order to constrain system parameters and look for transit timing and depth variations .
they also estimated the system parameters for tres-3b and found consistency with previous estimates .
their analysis of the transit timing data show no evidence for transit timing variations and timing measurements are able to rule out super - earth and gas giant companions in low order mean motion resonance with tres-3b .
the main aims of this study can be summarized in the following items : ( i ) determination of the system parameters for tres-3b ( two independent codes will be used ) and comparison with previous studies ( i.e. @xcite ) .
( ii ) based on the obtained transits , we will determine the mid - transit times ( @xmath0 ) and with following analysis of transit time variation ( ttv ) we will discuss possible presence of a hypothetical additional planet ( perturber ) .
we will try to estimate its upper - mass limit as a function of orbital periods ratio of transiting planet and the hypothetical perturber .
( iii ) finally , using the long - term integration and applying the method of maximum eccentricity we will search for stability of regions inside the tres-3b planet in context of additional planet(s ) .
the remainder of this paper is organized as follows . in the section 2 ,
we describe observations and data reduction pipelines used to produce the light curves .
section 3 presents the methods for analysis of transit light curves as well as discussion and comparison of the parameters of tres-3 system .
section 4 and 5 are devoted to ttv and long - term stability of the system , respectively .
finally , in section 6 we summarize and discuss our results .
we obtained our data using several telescopes with different instruments .
this allowed us to obtain many light curves since this strategy can effectively cope with the weather problems . on the other hand
, this approach results in rather heterogeneous data .
we used most of the data in average quality for the ttv analysis which is not very demanding on homogeneity of the data .
only the best quality light curve was used for the planet parameter determination .
observations used in this paper were carried out at the several observatories in slovakia ( star lesn observatory ; 4909 10``n , 2017 28''e ) , poland ( piwnice observatory ; 5305 43``n , 1813 46''e ) , germany ( grossschwabhausen observatory ; 5055 44``n , 1129 03''e ; volkssternwarte kirchheim observatory , 505544``n , 1129 03''e and michael adrian observatory , 495527``n , 0824 33''e ) and spain ( calar alto observatory ; 371325``n , 0232 46''e ) .
we collected 14 transit light curves obtained between may 2009 and september 2011 .
the transits on may 12 , 2009 and august 20 , 2010 were observed simultaneously at two different observatories .
the telescope diameters of 0.5 to 2.2 m allowed us to obtain photometry with 1.2 7.8 mmag precision , depending on observing conditions .
observations generally started @xmath15 1 hour before the expected beginning of a transit and ended @xmath15 1 hour after the event .
unfortunately , weather conditions and schedule constraints meant that we were not able to fit this scheme in all cases .
all instruments are equipped with ccd cameras with the johnson - cousins ( @xmath16 ) standard filter system . the information from individual observatories and instruments as well as the summary of observing runs
are given in table [ tab01 ] and table [ tab02 ] .
the standard correction procedure ( bias , dark and flat field correction ) and subsequently aperture photometry was performed by iraf and task _ chphot _
@xcite ( gsh and vk ) , c - munipack package ( g1 ) and mira_pro_7 ( ma ) .
data from remaining telescopes ( p and ca ) were reduced with the software pipeline developed for the semi
automatic variability search sky survey @xcite . to generate an artificial comparison star , at least 2030 per cent of stars with the lowest light
curve scatter were selected iteratively from the field stars brighter than 2.53 mag below the saturation level ( e.g. @xcite ) . to measure instrumental magnitudes ,
various aperture radii were used .
the aperture which was found to produce light curve with the smallest overall scatter was applied to generate final light curve .
the linear trend in the out - of - transit parts was also removed .
& telescope & detector & n@xmath17 + & & ccd size & fov [ arcmin ] + g1 & newton & sbig st10-mxe & 5 + & 508/2500 & 2184 @xmath18 1472 , 6.8 @xmath19 m & 20.4 @xmath18 13.8 + ma & cassegrain & sbig stl-6303e & 2 + & 1200/9600 & 3072 @xmath18 2048 , 9 @xmath19 m & 10 @xmath18 7 + gsh & ctk & site tk1024 & 1 + & 250/2250 & 1024 @xmath18 1024 , 24 @xmath19 m & 37.7 @xmath18 37.7 + & stk & e2v ccd42 - 10 & 1 + & 600/1758 & 2048 @xmath18 2048 , 13.5 @xmath19 m & 52.8 @xmath18 52.8 + vk & rct & stl-6303e & 2 + & 600/1800 & 3072 @xmath18 2048 , 9 @xmath19 m & 71 @xmath18 52 + p & cassegrain & sbig stl-1001 & 1 + & 600/13500 & 1024 @xmath18 1024 , 24@xmath19 m & 11.8 @xmath18 11.8 + ca & rcf & site ccd & 2 + & 2200/17037 & 2048 @xmath18 2048 , 24@xmath19 m & 18.1 @xmath18 18.1 + lcccc obs . & date & filter & @xmath20 & @xmath21 ( s ) + g1 & 2009 may 12 & @xmath22 & 319 & 40 + & 2009 aug 01 & @xmath22 & 345 & 45 + & 2010 apr 27 & @xmath22 & 180 & 35 + & 2010 jun 30 & @xmath22 & 238 & 40 + & 2010 aug 07 & @xmath22 & 168 & 35 + ma & 2010 july 13 & @xmath22 & 349 & 25 + & 2010 aug 20 & @xmath22 & 296 & 20 + gsh ( ctk ) & 2009 may 25 & @xmath23 & 138 & 80 + ( stk ) & 2011 mar 22 & @xmath22 & 131 & 60 + vk & 2009 aug 14 & @xmath24 & 103 & 120 + & 2010 aug 20 & @xmath25&320 & 30 + p & 2009 may 12 & @xmath22 & 120 & 55 + ca & 2010 sep 06 & @xmath22 & 158 & 30 - 35 + & 2011 sep 12 & @xmath22 & 128 & 45 +
our photometric observations of the tres-3 system consist of data from different instruments and are of different photometric quality . for the purpose of radius determination
, we decided to analyse only a light curve with the lowest scatter and the best rms = 1.2 mmag .
we choose data obtained at calar alto observatory on september 06 , 2010 ( see figs . 1 and 2 ) .
we first refined the light - curve system parameters and subsequently determined the individual times of transit as described below in order to improve ephemeris .
for the calculation of synthetic light curves we used two independent approaches : the first one is based on the routines from @xcite and the monte carlo simulations ( described in section [ sol1 ] ) , the second one uses jktebop code @xcite ( see section [ jktebop ] ) .
first we used the downhill simplex minimization procedure ( implemented in routine amoeba , @xcite ) to determine @xmath26 system parameters @xmath27 ( planet to star radius ratio ) , @xmath28 ( inclination ) , @xmath0 ( mid - transit time ) and @xmath29 ( star radius to semi - major axis ratio ) .
the model light curve itself was computed via the analytic expressions from @xcite .
the quadratic limb darkening law was assumed and corresponding limb darkening coefficients @xmath30 , @xmath31 were linearly interpolated from @xcite assuming the stellar parameters from @xcite : @xmath32 = 5650 k , @xmath33 ( g ) = 4.4 and [ fe / h ] = @xmath20.19 . as a goodness of the fit estimator we used the @xmath34 function : @xmath35 where @xmath36 is the model value and @xmath37 is the measured value of the flux , @xmath38 is the uncertainty of the @xmath39 measurement and the sum is taken over all measurements . the orbital period and the limb darkening coefficients were fixed through the minimization procedure .
the transit duration @xmath40 was determined assuming the semi - major axis @xmath41 = 0.02282@xmath42au @xcite .
the parameters and the corresponding light curve for which we found the minimum value of @xmath43 function , are in table 3 and figure 2 , respectively ( _ solution 1 _ ) . to estimate the uncertainties of the calculated transit parameters , we employed the monte carlo simulation method @xcite .
we produced 10000 synthetic data sets with the same probability distribution as the residuals of the fit in figure [ fig02 ] . from each synthetic data
set obtained by this way we estimated the synthetic transit parameters .
the minimum @xmath44 value corresponding to each set of monte carlo parameters was calculated as : @xmath45 where @xmath36 is the original best fit model value and @xmath46 is the monte carlo simulated value .
figure [ fig03 ] shows the dependence of the parameters @xmath27 and @xmath28 on the reduced @xmath47 .
this quantity is defined as : @xmath48 where @xmath49 is the number of data points and @xmath50 is the number of fitted parameters . to fully understand the errors of the system parameters we constructed confidence intervals in 2-dimensional space .
figure [ fig04 ] depicts the confidence regions for 2 parameters ( @xmath27 vs. @xmath28 ) as a projection of the original 4-dimensional region .
the gray - colored data points stand for the @xmath51 , @xmath52 and @xmath53 region with corresponding value of @xmath54 , @xmath55 and @xmath56 , respectively ( @xmath57 is the @xmath34 of the original best fit model value ) . from the shape of the dependence
it could be seen that these two parameters correlate ( see section [ analysis ] ) .
finally , we took into account the uncertainty of the stellar mass and semi - major axis according to the simple error propagation rule . the results from the first analysis
are shown in the table [ tab03 ] as solution 1 .
we have modelled the light curve using the jktebop code @xcite as well .
jktebop grew out of the original ebop program written for eclipsing binary systems @xcite via implementing the nde @xcite model .
jktebop uses biaxial spheroids to model the component stars ( or star and planet ) and performs a numerical integration in concentric annuli over the surface of each body to obtain the flux coming from the system .
this feature of the code allows us to avoid small and spherical planet approximations which are used in analytic light - curve generators based on @xcite , and hence to derive planet s oblateness .
a model is fitted to the data by the levenberg - marquardt least - square procedure .
the code converges rapidly toward a reliable solution and diminishes the correlation between fitted parameters @xcite .
, width=321 ] the main parameters of a jktebop fit are the orbital inclination @xmath28 , and the fractional radii of the host star and planet , @xmath58 and @xmath59 .
the fractional radii are defined as : @xmath60 where @xmath61 and @xmath62 are the stellar and planetary radii and @xmath41 is the orbital semi - major axis .
parameters @xmath58 and @xmath59 correspond to radii of spheres of the same volume as the biaxial spheroids . in jktebop
the fractional radii are reparametrized as their sum and ratio : @xmath63 because these are only weakly correlated with each other @xcite .
the directly fitted orbital inclination , @xmath28 , allows the transit impact parameter @xmath64 = @xmath65 cos @xmath28 to be calculated .
the initial values of parameters listed above were taken from @xcite .
because of different quality of data , the synthetic light curve was calculated only for the best light curve obtained at calar alto on september 06 , 2010 ( the same like in solution 1 ) .
a value of @xmath41 = 0.02282@xmath66 au @xcite was used in subsequent calculations .
the best - fit model was used as a template and fitted to other light curves for which only the mid - transit time was allowed to vary .
the resulting mid - transit times together with @xmath67 obtained from literature @xcite are analysed in details in the section 4 .
the determined parameters obtained by jktebop code are presented in table [ tab03 ] as solution 2 .
the errors of derived parameters were determined in two ways for each combination of data set and adopted ld law ( we have used the same coefficients like in the case of solution 1 ) .
firstly , we ran 1000 monte carlo ( mc ) simulations , a spread range of a given parameter within 68.3% was taken as its error estimate . secondly , the prayer - bead method ( e.g. @xcite ) was used to check whether red noise was present in our data .
mc errors were found to be 2 3 times smaller than the values returned by the prayer bead method .
this indicates that the light curve is affected not only by poisson noise but also by additional correlated noise .
therefore , our prayer - bead error estimates were taken as our final errors .
lccccc source & @xmath68 & @xmath69 & @xmath28 & @xmath70 & @xmath71 + & & & @xmath72 $ ] & [ min ] & [ days ] + @xcite & 0.1650 @xmath3 0.0027 & 0.1660 @xmath3 0.0024 & 82.15 @xmath3 0.21 & &
1.30619(1 ) + @xcite & 0.1687@xmath73 & 0.1655 @xmath3 0.0020 & 81.85 @xmath3 0.16 & &
1.30618581(51 ) + @xcite & & 0.1664@xmath74 & 81.73@xmath75 & 79.92@xmath76 & 1.3061864(5 ) + @xcite & & 0.1662@xmath77 & & 83.77@xmath78 & + @xcite & & & 82.07 @xmath3 0.17 & & 1.3061864(5 ) + @xcite & & 0.1603 @xmath3 0.0042 & 81.77 @xmath3 0.14 & &
1.30618700(15 ) + @xcite & 0.1664 @xmath3 0.0204 & 0.1661 @xmath3 0.0343 & 81.99 @xmath3 0.30 & 81.9 @xmath3 1.1 & 1.30618608(38 ) + @xcite & & & 81.93 @xmath3 0.13 & & 1.30618700(72 ) + @xcite & & & & 77.9 @xmath3 1.9 & 1.3061865(2 ) + @xcite & & & & & + solution_1 & 0.1675 @xmath3 0.0008 & 0.1652 @xmath3 0.0009 & 81.95 @xmath3 0.06 & & 1.3062132(2 ) + @xcite & & & & & + solution_2 & 0.1698 @xmath3 0.0014 & 0.1649 @xmath3 0.0015 & 81.51 @xmath3 0.14 & & 1.3062128(2 ) + @xcite & 0.1721@xmath79 & 0.1693@xmath80 & 81.35@xmath81 & 81.30 @xmath3 0.23 & 1.306 1854(1 ) + this work & & & & & + solution_1 & 0.1682 @xmath3 0.0032 & 0.1644 @xmath3 0.0047 & 81.86 @xmath3 0.28&79.20 @xmath3 1.38 & 1.306186 + this work & & & & & + solution_2 & 0.1696@xmath82 & 0.1669@xmath83 & 81.76@xmath84 & 79.08 @xmath3 0.72 & 1.306186 + the resulting values of the parameters together with their uncertainties are given in table [ tab03 ] .
for comparison , the parameters from previous studies are added there as well .
figure [ fig02 ] shows the resulting best fit obtained by solution 1 . in order to compare both solutions , we also plotted differences from the fit obtained by routines resulting from solution 1 and solution 2 .
the final parameters are in good agreement with already published values .
the output from solution 1 , in particular @xmath62 and @xmath61 , correspond a bit better to parameters of @xcite and @xcite .
the radii determined by jktebop are also within the range of errors .
based on these two parameters we also inferred the critical value of inclination for total transit of tres-3b as : @xmath85 as listed in table [ tab03 ] , all values of inclination including uncertainties are in agreement with critical inclination calculated above .
this is an evidence of grazing transit of tres-3b .
we used mc simulations to produce of 10000 synthetic data sets with the same probability distribution as the residuals of the fit . from each synthetic data
set we estimated the synthetic transit parameters . using the results of the simulation , the dependence of the inclination versus the planet to star radius ratio is plotted in figure [ fig04 ] .
the figure demonstrates that there is a correlation between the parameters and that solution for the system tres-3 is not unique and can be located in a relatively wide range ( degeneracy of the parameters ) .
this fact is also caused by grazing transit and by subsequent sensitivity of the solution to the determination of ld coefficients . for determination of our mid - transit times , the best - fit model was used as a template and fitted to other light curves for which only the mid - transit time was allowed to vary . in order to have homogeneous analysis of ttv , we collected the data points of transit light curves published in previous studies ( see figure [ fig05 ] ) .
transit light curve data of @xcite and @xcite were available on - line and the remaining tabulated data were obtained from mentioned authors by e - mail .
we re - analysed all transit light curves and derived mid - transit times by the same procedure ( the code jktebop ) to get a homogeneous dataset for ttv analysis ( see section [ ttv ] ) .
we also re - analysed all light curves under consideration with @xmath86 as a free parameter and tried to search for any variation .
all determinations of @xmath87 were consistent within error bars with a mean value , so we did not detect any transit depth variation .
we also saw no significant signal in periodograms .
our new 14 mid - transit times and 42 redetermined literature values were used to refine the transit ephemeris .
the mid - transit times were transformed from jd or hjd ( based on utc ) into bjd ( based on barycentric dynamical time
tdb ) using the on - line converter by @xcite . as a result of fitting a linear function of the epoch
, we obtained the mid - transit time for the initial epoch @xmath88 bjd@xmath89 and the orbital period @xmath90 d. the individual mid - transit errors were taken as weights .
the linear fit yields reduced @xmath43 of @xmath91 that is similar to a value of 2.3 reported by @xcite .
these values , noticeable greater than 1 might suggest the existence of an additional planet which perturbs the orbital motion of tres-3 b @xcite . lcclll date & obs . & epoch & @xmath92 @xmath93 & @xmath92 error & @xmath94 ( d ) + 2009 may 12 & p & 326 & 2454964.39885 & 0.00084 & @xmath95 + 2009 may 12 & g1 & 326 & 2454964.39859 & 0.00081 & @xmath96 + 2009 may 25 & gsh ( ctk ) & 336 & 2454977.4593 & 0.0016 & @xmath97 + 2009 aug 01 & g1 & 388 & 2455045.38091 & 0.00045 & @xmath98 + 2009 aug 14 & vk & 398 & 2455058.44418 & 0.00077 & @xmath99 + 2010 apr 27 & g1 & 594 & 2455314.45510 & 0.00062 & @xmath100 + 2010 jun 30 & g1 & 643 & 2455378.45937 & 0.00031 & @xmath101 + 2010 july 13 & ma & 653 & 2455391.52067 & 0.00017 & @xmath102 + 2010 aug 07 & g1 & 672 & 2455416.3397 & 0.0011 & @xmath103 + 2010 aug 20 & ma & 682 & 2455429.40118 & 0.00033 & @xmath104 + 2010 aug 20 & vk & 682 & 2455429.3997 & 0.0004 & @xmath105 + 2010 sep 06 &
ca & 695 & 2455446.38075 & 0.00017 & @xmath106 + 2011 mar 22 & gsh ( stk ) & 846 & 2455643.6145 & 0.0005 & @xmath107 + 2011 sep 12 & ca & 979 & 2455817.3372 & 0.0014 & @xmath107 + results for new mid - transit times are shown in table [ tab04 ] .
the observed minus calculated ( o c ) diagram , plotted in fig . [ fig05 ] , shows no significant deviation from the linear ephemeris .
all our data points deviated by less than @xmath108 .
we also searched for a periodicity that could be a sign of an additional body in the system .
we generated a lomb
scargle periodogram @xcite for the residuals in a frequency domain limited by the nyquist frequency and found the highest peak at @xmath1530 d and peak - to - peak amplitude of @xmath109 s. this period could coincide with a stellar rotation period which is roughly estimated to be @xmath1528 d. examples of such ttv signals induced by the stellar activity are observed in the kepler data ( e.g. @xcite ) .
however , the false alarm probability ( calculated empirically by a bootstrap resampling method with @xmath110 trials ) of the putative signal is disqualifying with a value of 18.2% .
in addition , the amplitude is close to the mean 1-@xmath111 timing error of our observations .
these findings allow us to conclude that a strictly periodic ttv signal with the amplitude greater than @xmath151 minute over a 4-year time span seems to be unlikely . following @xcite , we put upper constraints on a mass of a potential perturbing planet in the system with refined assumptions .
we adopted a value of 1 min for the maximal amplitude of the possible ttv signal .
we also increased the sampling resolution to probe resonances other than inner and outer 1:2 commensurabilities studied in detail by @xcite .
we simplified the three - body problem by assuming that the planetary system is coplanar and initial orbits of both planets are circular .
the masses of both the star and the transiting planet , as well as its semi - major axis were taken from @xcite .
the orbital period of the hypothetical perturber was varied in a range between 0.2 and 5 orbital periods of tres-3 b. we produced 2800 synthetic o
c diagrams based on calculations done with the mercury code @xcite .
we applied the bulirsch
stoer algorithm to integrate the equations of motion . , and the hypothetical perturber , @xmath112 .
orbits located in a grey area were found to be unstable due to close encounters of both planets.,width=321 ] the most important feature of the bulirsch - stoer algorithm for n - body simulations is that it is capable of keeping an upper limit on the local errors introduced due to taking finite time - steps by adaptively reducing the step size when interactions between the particles increase in strength .
calculations covered 1500 days , i.e. a total time span of available transit observations . the results of simulations are presented in fig .
[ fig06 ] .
our analysis allows us to exclude an additional earth - mass planet close to inner 1:3 , 1:2 , and 3:5 and outer 5:3 , 2:1 , and 3:1 mean - motion resonances ( mmrs ) .
in this section , we investigated the long - term gravitational influence of tres-3b planet on a potential second planet in the system .
thus , we performed numerical simulation for studying the stability of orbits and checking their chaotic behavior using the method of maximum eccentricity ( e.g. @xcite ) .
we used long - term integration of small - mass ( earth - mass ) planet orbits for inspecting the stability regions in the tres-3b system .
the time span of the integration was around 140000 revolutions of the planet around the star .
for the integration of this system we again applied an efficient variable - time - step algorithm ( bulirsch - stoer integration method ) .
the parameter @xmath113 which controls the accuracy of the integration was set to @xmath114 , in this case .
we have generated @xmath110 mass - less particles for representing small planets in this system .
we assumed that semi - major axis ranges from 0.00625@xmath1au to 0.1@xmath1au .
the inner border of the generated system of small planets is around 1.5 times star radius .
this value of inner border is approximately at the location of roche limit for a earth - like planet and includes also the region in the vicinity of 3:1 mmr ( 0.01au for this case ) in our analysis .
the step - size in semi - major axis was @xmath115 = 0.00025@xmath1au , in eccentricity @xmath116 = 0.02 and @xmath117 = 5@xmath118 in inclination .
the upper limit of the grid in eccentricity was 0.5 and 50@xmath118 in inclination .
we integrated the orbits of the small planets for the time - span of 500 years ( about 140 000 revolutions of tres-3b around the parent star ) .
we obtained orbital evolution of each small planet in the system and also were able to study the stability regions at the end of the integration using the method of maximum eccentricity , mentioned above . the maximum eccentricity that the potential perturber of tres-3b reached over the time of the integration
is plotted on the @xmath119 ( @xmath120 ) stability map in the fig.[fig_ae ] ( fig.[fig_ai ] ) .
figure [ fig_ae ] shows the stability maps in @xmath119 plane for selected values of inclinations : @xmath121 , @xmath122 , @xmath123 and @xmath124 ( from top left to bottom right ) .
also the mmrs with tres-3b planet are marked in these plots .
we can see a stable region inside the 2:1 mmr for each value of inclination . outside the 2:1
mmr the gravitational influence of the tres-3b planet is very strong and leads to depletion of these regions .
only the mmrs 2:1 , 3:2 and 1:1 are moderately populated but the population decreases with the increase of the inclination . for completeness , we note that 5:3 ( @xmath125 0.016au ) and 3:5 ( @xmath125 0.032au ) mmrs are depleted and thus not stable for additional earth - mass planet in their vicinity . in fig.[fig_ai ] we present the stability maps in @xmath120 plane for several values of eccentricities : @xmath126 , @xmath127 , @xmath128 and @xmath129 .
one can see the stable region inside the 2:1 mmr with tres-3b planet , which we found stable also in the @xmath119 plane .
considering the region beyond the 1:3 mmr , the depletion of the planet population is not so strong than in the region between 2:1 and 1:3 mmr .
this feature is in a good agreement with the weaker gravitational influence of the tres-3b planet .
based on our results , we showed that the region inside the tres-3b planet orbit ( especially inside the 2:1 up to 3:1 mmr ) can be stable on longer timescales ( hundreds of years or hundred thousands of revolutions of tres-3b around the star ) from the dynamical point of view .
the region from 0.015@xmath1au ( near 2:1 mmr ) to 0.05@xmath1au ( near 1:3 mmr ) was found unstable apart from moderately populated mmrs located in this area ( see figures fig.[fig_ae ] and fig.[fig_ai ] ) .
the relatively small increase of population in the mentioned mmrs depends on the initial values of semi - major axis , eccentricity and inclination .
the region beyond the 0.05@xmath1au was found to have a chaotic behavior and the depletion of the planet population increases with increasing values of initial eccentricity as well as inclination .
based on the transit light curves obtained at several observatories between may 2009 and september 2011 we redetermined orbital parameters and the radius of the transiting planet tres-3b .
the best light curve ( obtained at calar alto , sep 2010 ) was used for light curve analysis , and the data from the other observatories were used for @xmath0 determination and ttv investigation .
we used two independent solutions for parameters determination and finally , we concluded that our values are consistent with previous results of @xcite and @xcite .
the aim of this present paper was also to discuss possible presence of a second planet in the extrasolar system tres-3 .
for this purpose we used our new 14 mid - transit times and the individual determinations from @xcite and @xcite .
the resulting @xmath94 diagram showed no significant deviation of data points from the linear ephemeris .
in addition , we tried to search for a periodicity that could be caused by an additional body in the system .
we can conclude that a strictly periodic ttv signal with the amplitude greater than 1 minute over a 4-year time span seems to be unlikely .
this result , together with refined assumptions of @xcite allow us to put upper constraints on the mass of a potential perturbing planet .
the additional earth - mass planet close to inner 3:1 , 2:1 , and 5:3 and outer 3:5 , 1:2 , and 1:3 mmrs can be excluded .
finally , we used the long - term integration of the theoretical set of massless particles generated in tres-3 system for studying the dynamical stability of potential second planet in the system ( influenced by tres-3b gravitation ) . from our analysis
we found that the region inside the tres-3b planet orbit ( especially inside the 2:1 mmr ) up to 3:1 mmr can be stable on longer timescales ( hundreds of years or hundred thousands of revolutions of tres-3b around the star ) from the dynamical point of view .
the region from 0.015@xmath1au ( near 2:1 mmr ) to 0.05@xmath1au ( near 1:3 mmr ) was found to be unstable apart from moderately populated mmrs located in this area .
the relatively small increase of population in these mmrs depends on the initial values of semi - major axis , eccentricity and inclination .
the region beyond 0.05@xmath1au was found to have a chaotic behavior and depletion of the planet population increases with increasing values of initial eccentricity as well as inclination .
this work has been supported by the vega grants no .
2/0078/10 , 2/0094/11 , 2/0011/10 . mv , mj , jb , tp and p would like to thank also the project apvv-0158 - 11 . tk thanks the student project grant at mu muni / a/0968/2009 and the national scholarship programme of the slovak republic .
gm acknowledges iuventus plus grants ip2010 023070 and ip2011 031971 .
sr thanks the german national science foundation ( dfg ) for support in project ne 515 / 33 - 1 .
broeg c. , fernndez m. , neuhuser r. 2005 , an , 326 , 134 chambers j.e .
1999 , mnras , 304 , 793 charbonneau d. , brown t. m. , latham d. w. , mayor m. 2000 , apj , 529 , l45 christiansen j. l. , ballard s. , charbonneau d. 2011 , apj , 726 , 94 claret a. 2000 , a&a , 363 , 1081 collier cameron a. , wilson d. m. , west r. g. , et al .
2007 , mnras , 380 , 1230 coln k. d. , ford e. b. , lee b. , et al .
2010 , mnras , 408 , 1494 dsert j. m. , sing d. , vidal - madjar a. , et al .
2011 , a&a , 526 , 12 dvorak , r. , pilat - lohinger , e. , funk , b. , freistetter , f. 2003 , a&a , 398 , l1 eastman j. , siverd r. , gaudi b.s .
2010 , pasp , 122 , 935 etzel p. b. 1981 , in carling e. b. , kopal z. , eds , nato asi ser .
c. , 69 , photometric and spectroscopic binary systems , kluwer , dordrecht , 111 gibson n. p. , pollacco d. , simpson e. k. , et al .
2009 , apj , 700 , 1078 kundurthy p. , becker a. c. , agol e. , et al .
2013 , apj , 764 , 8 lee j. w. , youn j. h. , kim s. l. , et al .
2011 , pasj , 63 , 301 lomb n.r . 1976 , ap&ss , 39 , 447 mazeh t. , nachmani g. , holczer t. , et al .
2013 , arxiv:1301.5499v1 mandel k. , agol e. 2002 , apj , 580 , l171 mugrauer m. 2009 , an , 330 , 419 mugrauer m. , berthold t. 2010 , an , 331 , 449 nelson b. , davis w. d. 1972 , apj , 174 , 617 niedzielski a. , maciejewski g. , czart k. 2003 , acta astronomica , 53 , 281 odonovan f. t. , charbonneau d. , bakos g. . , et al .
2007 , apj , 663 , l37 popper d. m. , etzel p. b. 1981 , aj , 86 , 102 press w. h. , teukolsky s. a. , vetterling w. t. , flannery b. p. 1992
, _ numerical recipes in _ fortran , cambridge : university press , london raetz s. , mugrauer m. , schmidt t. o. b. , et al .
2009 , an , 330 , 475 sada p. v. , deming , d. jennings d. e. , et al .
2012 , pasp , 124 , 212 scargle j.d . 1982 , apj , 263 , 835 southworth j. , maxted p. f. l. , smalley b. 2004 , mnras , 349 , 547 southworth j. 2008 , mnras , 386 , 1644 southworth j. 2010 , mnras , 408 , 1689 southworth j. 2011 , mnras , 417 , 2166 southworth j. 2012 , mnras , 426 , 1291 sozzetti a. , torres g. , charbonneau d. , et al .
2009 , apj , 691 , 1145 torres g. , winn j. n. , holman m. 2008 , apj , 677 , 1324 turner j. d. , smart b. m. , hardegree - ullman k. k. et al .
2013 , mnras , 428 , 678 winn j. n. , holman m. j. , henry g. w. , et al .
2009 , apj , 693 , 794 | we present new observations of the transiting system tres-3 obtained from 2009 to 2011 at several observatories .
the orbital parameters of the system were re- determined and a new linear ephemeris was calculated .
the best quality light curve was used for light curve analysis , and other datasets were used to determine mid - transit times ( @xmath0 ) and study transit time variation ( ttv ) . for planet parameter determination we used two independent codes and finally , we concluded that our parameters are in agreement with previous studies .
based on our observations , we determined 14 mid - transit times . together with published @xmath0 we found that the timing residuals showed no significant deviation from the linear ephemeris .
we concluded that a periodic ttv signal with an amplitude greater than 1 minute over a 4-year time span seems to be unlikely .
our analysis of an upper mass limit allows us to exclude an additional earth - mass planet close to inner 3:1 , 2:1 , and 5:3 and outer 3:5 , 1:2 , and 1:3 mean - motion resonances . using the long - term integration and applying the method of maximum eccentricity , the region from 0.015@xmath1au to 0.05@xmath1au
was found unstable and the region beyond the 0.05@xmath1au was found to have a chaotic behaviour and its depletion increases with increasing values of the initial eccentricity as well as inclination .
[ firstpage ] planets and satellites : individual : tres-3b stars : individual : tres-3 |
the presence of extended x - ray emitting coronae associated with present - day massive disc galaxies is a generic prediction of cold dark matter ( cdm ) galaxy formation theory ( e.g. , @xcite ) .
the coronae serve as a reservoir of both diffuse , metal - poor gas accreted from the intergalactic medium ( igm ) , and metal - rich gas that is either ejected from galaxies by energetic feedback or stripped from infalling satellites .
two mechanisms are thought to be chiefly responsible for establishing the the x - ray emission properties of coronae : feedback injected by stellar wind / supernovae ( sne ) ( e.g. , @xcite ) and active galactic nuclei ( agn ) , and the shock heating and adiabatic compression of gas accreted from the igm ( e.g. , @xcite ) .
although galactic coronae have been detected around many disc galaxies , their origin remains uncertain .
for example , the coronal luminosity is observed to correlate well with many tracers of feedback activity [ e.g. , the infrared ( ir ) or radio luminosities , @xcite ] , and the heavy element abundance of the x - ray luminous gas is also indicative of enrichment by sne ( e.g. , @xcite ) .
these are often regarded as direct evidence for the feedback scenario ( but see @xcite for an alternative interpretation ) .
furthermore , non - gravitational heating is also indicated by the change of slope of the scaling relations between the coronal luminosity / temperature and the galaxy mass in different mass ranges ( e.g. , @xcite ) .
the uncertainty is driven by the difficulty of detecting and characterizing the diffuse thermal x - ray emission at large galactocentric radii ( e.g. , @xcite ) . in recent years , with either stacking of large numbers of x - ray observations of various types of galaxies @xcite or moderately deep single - pointing x - ray observations of massive disc galaxies @xcite , x - ray emission from the putative coronae on large scales ( at radii up to @xmath7 ) has been reported around several galaxies which are not particularly active in sf .
these coronae are considered to be most likely produced by the accretion of intergalactic gas . however
, such detections are still rare and only concern galaxies within a narrow range of stellar mass ( see table [ table : galaxypara ] ) .
comparison with x - ray measurements of galaxies in a broad mass range is thus particularly desirable . in our previous studies @xcite ( hereafter , paper i , ii ) , we investigated the galactic coronae associated with a sample of _ chandra_-observed nearby highly - inclined disc galaxies .
we found a tight correlation between the coronal luminosity and the sn energy input rate ; the abundance ratio of the coronal gas is also consistent with a combined contribution from core collapsed and type ia sne .
these results can be qualitatively interpreted , naively , as evidence that galactic coronae are mainly established by sn feedback .
however , our sample only includes low and intermediate mass galaxies ( typically with stellar mass @xmath8 ) , which may not be massive enough for accreted gas to contribute a dominant fraction of the detected x - ray flux .
in fact , recent cosmological hydrodynamical simulations invoking both accretion and feedback predict a coronal luminosity range that is broadly consistent with x - ray observations @xcite .
these simulations trace the dynamics of the coronal gas , a large fraction of which is quasi - hydrostatic or even infalling . @xcite further showed that the high ( @xmath9solar ) metallicities inferred from the x - ray spectroscopy are not in conflict with models in which the bulk ( by mass ) of the hot circumgalactic medium ( cgm ) is established by accretion .
the soft x - ray emission is dominated by collisionally - excited metal ions deposited by feedback , whilst the majority of the accreted gas is metal poor and is comparatively radiatively inefficient .
the hot cgm thus appears to be typically metal - poor in a mass - weighted sense , but its x - ray luminosity - weighted metallicity is often close to solar .
therefore , accretion of intergalactic gas may still be important , especially in those massive galaxies not included in our _ chandra _ sample . in paper
ii , we also compared our measurements to x - ray observations of massive elliptical galaxies , which have significantly steeper @xmath10 relations .
the formation of such galaxies may , however , be subject to more violent mergers and the corresponding starbursting process , potentially complicating the interpretation of observations of their coronae .
furthermore , the typically rich clustered environment of massive elliptical galaxies also potentially contaminates the measurement of coronal properties associated with individual galaxies .
therefore , comparisons with x - ray measurements of elliptical galaxies do not offer the cleanest means by which to place direct constraint on disc galaxy formation models
. we herein conduct a quantitative comparison of our x - ray measurements of disc galaxies to the x - ray measurements of several isolated massive disc galaxies from the literature ( [ paperiiisubsec : massivegalaxy ] ) , as well as the results of hydrodynamical simulations invoking both accretion and feedback ( i.e. , the gimic simulations , see [ paperiiisubsec : simulations ] ) , in order to further explore the origin of coronae around disc galaxies and place direct constraint on galaxy formation models .
this paper is organized as follows . in
[ paperiiisec : data ] , we describe our procedure for correcting coronal luminosity measurements to ensure data homogenization between the galaxies selected from our original _
sample , galaxies from the literature , and the gimic simulations .
we compare the observational results to an analytical model as well as the gimic simulations in [ paperiiisec : comparison ] , and further discuss the scientific implications in [ paperiiisec : discussion ] .
our results and conclusions are summarized in [ paperiiisec : summary ] .
in order to compare measurements from different observations and simulations in a uniform fashion , corrections to ensure the x - ray measurements for data homogenization are necessary .
we aim to compare the coronal luminosity measured in a consistent manner , i.e. , in the same band and in the same radial range . here , we consider only field galaxies from our original _
chandra _ sample ( paper i ) , because x - ray measurements of the clustered galaxies are potentially contaminated by the intracluster medium , and also because the massive disc galaxies whose measurements we collate from the literature ( [ paperiiisubsec : massivegalaxy ] ) and the selected galaxies from the gimic simulations ( [ paperiiisubsec : simulations ] ) are in the field .
we also select galaxies with published rotation velocity measurements ( inclination - corrected , @xmath11 , as listed in table [ table : galaxypara ] ) , which are used for estimating the halo masses and radii ( see below ) .
one galaxy , ngc 3384 , has extremely low @xmath11 ( @xmath12 ) ; this is likely a result of interaction with its companions , and hence an inaccurate reflection of the depth of its gravitational potential ( paper i ) .
this galaxy is thus excluded from the current sample . in total , 30 galaxies are selected and their key parameters are summarized in table [ table : galaxypara ] .
lccccccccccc ic2560 & 10.03 & @xmath13 & @xmath14 & @xmath15 & 225 & @xmath16 + m82 & 10.30 & @xmath17 & @xmath18 & @xmath19 & 108 & @xmath20 + ngc0024 & 9.18 & @xmath21 & @xmath22 & @xmath23 & 101 & @xmath24 + ngc0520 & 10.57 & @xmath25 & @xmath26 & @xmath27 & 76 & @xmath28 + ngc0660 & 10.47 & @xmath29 & @xmath30 & @xmath31 & 157 & @xmath32 + ngc0891 & 10.69 & @xmath33 & @xmath34 & @xmath35 & 245 & @xmath36 + ngc1023 & 10.83 & - & @xmath37 & @xmath38 & 124 & @xmath39 + ngc1482 & 10.36 & @xmath40 & @xmath41 & @xmath42 & 133 & @xmath43 + ngc1808 & 10.58 & @xmath44 & @xmath45 & @xmath46 & 135 & @xmath47 + ngc2787 & 10.52 & @xmath48 & @xmath49 & @xmath50 & 207 & @xmath51 + ngc2841 & 10.99 & @xmath52 & @xmath53 & @xmath54 & 379 & @xmath55 + ngc3079 & 10.47 & @xmath56 & @xmath57 & @xmath58 & 242 & @xmath59 + ngc3115 & 10.83 & - & @xmath60 & @xmath61 & 117 & @xmath62 + ngc3198 & 10.02 & @xmath63 & @xmath64 & @xmath65 & 166 & @xmath66 + ngc3521 & 10.85 & @xmath67 & @xmath68 & @xmath69 & 271 & @xmath70 + ngc3556 & 10.21 & @xmath71 & @xmath72 & @xmath73 & 172 & @xmath74 + ngc3628 & 10.83 & @xmath75 & @xmath76 & @xmath77 & 248 & @xmath78 + ngc3955 & 10.11 & @xmath79 & 86 & 10.95 & 93 & @xmath80 + ngc4244 & 8.95 & @xmath81 & @xmath82 & @xmath83 & 96 & @xmath84 + ngc4594 & 11.19 & @xmath85 & @xmath86 & @xmath87 & 430 & @xmath88 + ngc4631 & 10.01 & @xmath89 & @xmath90 & @xmath91 & 155 & @xmath92 + ngc4666 & 10.61 & @xmath93 & @xmath94 & @xmath95 & 221 & @xmath96 + ngc5102 & 9.18 & @xmath97 & 90 & 11.01 & 97 & @xmath98 + ngc5170 & 10.67 & @xmath99 & 245 & 12.41 & 285 & @xmath100 + ngc5253 & 8.69 & @xmath101 & @xmath102 & @xmath103 & 38 & @xmath104 + ngc6503 & 9.50 & @xmath105 & @xmath106 & @xmath107 & 70 & @xmath108 + ngc6764 & 10.00 & @xmath109 & @xmath110 & @xmath111 & 156 & @xmath112 + ngc7090 & 9.17 & @xmath113 & 102 & 11.19 & 111 & @xmath114 + ngc7582 & 10.83 & @xmath115 & @xmath116 & @xmath15 & 223 & @xmath117 + ngc7814 & 10.85 & - & @xmath118 & @xmath119 & 268 & @xmath120 + ngc266 & 11.30 & 2.4 & @xmath121 & 12.90 & 410 & @xmath122 + ngc1961 & 11.62 & 15.5 & @xmath123 & 13.08 & 470 & @xmath124 + ngc6753 & 11.51 & 11.8 & @xmath125 & 13.00 & 440 & @xmath126 + ugc12591 & 11.91 & 4.8 & @xmath127 & 13.38 & 601 & @xmath128 [ table : galaxypara ] we first estimate the virial mass and radius of each dark matter halo .
we represent these quantities by @xmath2 and @xmath129 , respectively , which are the total mass and radius of the halo with a mean density of @xmath130 [ we use the hubble constant from the _ wmap _ 9-years data @xcite to calculate @xmath131 , which is the critical density of the universe ] .
we estimate @xmath2 using the scaling relation applicable to dark matter halos from @xcite ( @xmath132 ) .
@xmath129 is computed using the definition of @xmath2 : @xmath133 . @xmath2 and @xmath129 of the sample galaxies
are listed in table [ table : galaxypara ] . for consistency , we measure the coronal properties within a radial range that scales with the galaxy mass or @xmath129 . the lower limit of the range should be large enough to exclude the x - ray emission from the galactic disc , while the upper limit should match or not significantly exceed the scale accessed by the _ chandra _ observations . for the typical coronal size of our _ chandra _ sample ( paper i ) , we choose this range to be @xmath134 .
this range also covers a significant fraction of the x - ray emission of coronae in hydrodynamical simulations ( e.g. , @xcite ) .
we renormalize the coronal luminosity measured within a typical vertical range of @xmath135 ( @xmath136 is the exponential scale height of the 0.5 - 1.5 kev diffuse x - ray intensity profile ) and radial range of @xmath137 ( @xmath138 is the b - band diameter of the projected major axis at the isophotal level of @xmath139 ) by extrapolating the intensity profile with an exponential model characterized by the diffuse x - ray vertical and radial scale heights of the sample galaxies ( paper i ) .
the exponential model used to fit the intensity profiles potentially inaccurately describes the x - ray intensity distribution at large radii , but the choice of @xmath140 as the upper limit of the radii range ( typically not exceeding @xmath135 and @xmath137 ) minimizes the uncertainty resulting from inaccurate exponential fits .
the renormalized coronal luminosities of the sample galaxies are listed in table [ table : galaxypara ] , with the original statistical errors from the coronal luminosity measurements ( paper i ) .
our _ chandra _ survey does not include galaxies with @xmath141 .
these massive disc galaxies are rare in nature , and consequently no such galaxy satisfies the selection criteria established in paper i. several recent studies , however , report the detection of large - scale diffuse x - ray emission associated with massive disc galaxies in the local universe ( e.g. , @xcite for _ chandra _ observations of ngc 1961 , @xcite for _ xmm - newton _ observations of ugc 12591 , @xcite for _ xmm - newton _ observations of ngc 1961 and ngc 6753 , and @xcite for _ rosat _ and _ chandra _ observations of ngc 266 ) .
we supplement our sample with these measurements , applying appropriate corrections to their coronal luminosities to ensure a uniform comparison .
@xmath2 and @xmath129 of ngc 1961 , ngc 6753 , and ugc 12591 are directly taken from the above cited references or estimated from the inclination - corrected rotation velocities , in the same way as for our _ chandra _ sample galaxies ( [ paperiiisubsec : oursample ] ) . however , the inclination angle of ngc 266 ( @xmath142 ) is too low to enable a reliable estimate of the rotation velocity ( the apparent maximum rotation velocity of gas is @xmath143 , c.f the inclination corrected @xmath144 ) .
we therefore use the halo mass estimated from the stellar mass of the galaxy , as adopted in @xcite . for ngc 266 , ngc 1961 ( for which the results from @xcite are used throughout the present paper ) , and ngc 6753 , the coronal luminosities are measured in the 0.5 - 2 kev band , same as for our _ chandra _ sample . for ugc 12591 ,
however , the original coronal luminosity from @xcite is measured in the 0.6 - 1.4 kev band .
we then adopt the spectral model described in @xcite ( @xmath145 , 0.5 solar abundance , and galactic foreground absorption column density ) to estimate the luminosity in the 0.5 - 2 kev band .
this correction factor is only of order @xmath146 .
we also adopt a redshift - independent distance of @xmath147 to ugc 12591 ( from the ned ) instead of the artificial 100 mpc distance taken in @xcite .
relevant galaxy parameters ( @xmath148 , sfr , and @xmath149 ) are also corrected for this revised distance estimate .
the largest correction factor is for the radial range adopted in measuring the coronal luminosity .
according to the cited references , this range is @xmath150 for ngc 266 @xcite , ngc 1961 , and ngc 6753 @xcite , but is within 50 kpc for ugc 12591 @xcite .
we use the model ( @xmath151-model or modified @xmath151-model ) adopted in fitting the radial intensity profiles as described in the references to renormalize the coronal luminosity to our adopted @xmath134 range .
however , the published _
rosat _ and _ chandra _ observations of ngc 266 are insufficiently deep for such a radial profile analysis .
we therefore use the ratio of the count rate within the @xmath134 and @xmath150 ranges ( after subtracting the contribution from the scattered photons of the central low luminosity agn ) to adopt a rough correction ( private communication with @xmath152 bogd@xmath153n ) .
we also subtract the estimated contributions from stellar sources [ bright low mass x - ray binaries ( lmxbs ) , @xcite ; faint lmxbs , cataclysmic variables ( cvs ) and coronal active binaries ( abs ) , @xcite ; high mass x - ray binaries ( hmxbs ) , @xcite ] within @xmath134 of ngc 266 , ngc 1961 , and ngc 6753 , by scaling from their stellar mass and sfr .
these stellar source contributions are not accounted for in the original references .
the final corrected coronal luminosities are listed in table [ table : galaxypara ] .
we use the results from the galaxies - intergalactic medium interaction calculation ( gimic ; @xcite ) , a suite of hydrodynamic resimulations of regions drawn from the millennium simulation @xcite .
the gimic simulations are performed with a variant of the treepm - sph code gadget3 , a substantial upgrade of gadget2 @xcite .
gimic follows the evolution of five representative roughly spherical regions with different overdensities drawn from the dark matter millennium simulation @xcite .
the gimic simulations have intermediate ( @xmath154 ) and high resolution ( @xmath155 ) , compared to the `` low '' resolution for the original millennium simulation , in which the collisionless particles , representing a composite of baryonic and dark matter , have mass of @xmath156 . in the present paper , we use the intermediate resolution simulations , all of which are run to z=0 .
@xcite demonstrate that the properties explored here are numerically converged in the regimes of interest , and that the resolution is sufficient to enable morphological classification of @xmath157 galaxies .
metals typically dominate the specific emissivity of astrophysical plasmas ( e.g. , @xcite ) .
therefore , the accurate prediction of coronal x - ray luminosities requires that simulations , besides modelling the gravitational and thermodynamical evolution of gas , must track the nucleosynthesis of metal species by stellar populations , and macroscopic transport of this material throughout the cosmological growth and assembly of galaxies ( and groups and clusters of galaxies ) .
these processes are modeled in gimic , and have been shown to play a key role in the establishment of the amplitude , and radial profile of , the luminosity and metallicity of coronae @xcite .
one shortcoming of gimic , however , is that it does not model the evolution of black holes or feedback effects associated with them .
therefore , we caution that the stellar mass and the radiative cooling ( and hence the coronal x - ray luminosity ) of the most massive galaxies in the simulated sample ( where agn feedback is expected to be non - negligible ) may be inadequately modeled ( see [ paperiiisubsec : agnfeedback ] for further discussion ) .
we adopt the same disc galaxy sample as @xcite , selecting central galaxies within friends - of - friends ( fof ) halos , and focussing on @xmath157 disc galaxies by requiring a disc - to - total stellar mass ratio @xmath158 and a stellar mass in the range @xmath159 .
@xcite considered only isolated systems by excluding galaxies that are interacting or are members of galaxy groups and clusters . as per @xcite ,
the x - ray luminosity of a corona is computed by summing the luminosities of all gas particles bound to the corresponding subhalo .
the cooling function is computed using the astrophysical plasma emission code ( apec , v1.3.1 , @xcite ) under the assumption that the gas is in collisional ionization equilibrium . as is generally the case for cosmological simulations ,
the adopted resolution is insufficient to model the formation of individual stars within a multi - phase interstellar medium .
the simulations therefore impose a polytropic equation of state to the high density ( @xmath160 ) gas that is subject to thermo - gravitational instability ( e.g. @xcite ) .
this gas ( mostly in the disc ) is assigned a temperature of that is below the minimum considered by the apec cooling tables ; therefore , by construction , the computed luminosity comprises only extraplanar emission .
we herein calculate the coronal luminosity within @xmath134 , which is dominated by hot gas and less affected by the multi - phase gas in the disc . to renormalize the original coronal x - ray luminosities from @xcite , which were measured for the entire halo to @xmath161 , we compute the average stacked radial intensity profiles in five halo mass bins [ @xmath162 , and then calculate the luminosity fraction enclosed within @xmath134 .
this fraction is typically @xmath163 of the total coronal luminosity .
we corrected this enclosed fraction to compute the 0.5 - 2 kev luminosity within @xmath134 from the original total coronal luminosity .
the sfr ( stellar mass ) of a simulated galaxy is estimated by summing the sfrs ( stellar masses ) of all gas ( star ) particles bound to a galaxy s subhalo .
the rotation velocity , @xmath11 , is obtained from the maximum value of the halo velocity curve , assuming spherical symmetry .
before comparing our x - ray measurements to results from gimic , it is instructive to consider a simple analytical model involving only gravitational energy release of the accreted gas ( no feedback ) as a reference for the comparison .
the inaccuracy of this simple accretion - only model @xcite , as compared to gimic , was highlighted in @xcite . in this subsection
, we adopt the uniformly corrected coronal luminosity from papers i and ii , without any further corrections as described in [ paperiiisubsec : oursample ] , in order to include the x - ray emission from the galactic disc .
quantitative comparisons with x - ray measurements of massive disc galaxies and the results from gimic will be presented in the following sections .
we estimate the expected coronal x - ray luminosity ( @xmath164 ) , in the absence of any effects of feedback on the mass , structure and emissivity of the hot cgm , following the procedure detailed in @xcite . assuming the accretion rate @xmath165 can be connected to the sfr by the poorly constrained sf efficiency ( @xmath166 ; e.g. , @xcite ) , @xmath164 can be predicted with the observational parameters @xmath11 and sfr as : @xmath167 assuming @xmath168 , we then compare @xmath164 to the observed x - ray luminosity from our measurements ( fig .
[ fig : analyticalmodel]a ) and obtain a best - fit linear relation of ( see 3.1 of paper ii for the method of fitting ) : @xmath169 the relation is just qualitative , showing the general correlation and large inconsistency between the gravitational energy and the x - ray luminosity .
the coefficient is , clearly , far below unity ( @xmath170 ) , consistent with conclusions reached previously ( e.g. , @xcite ) . reducing @xmath171 to a value significantly below unity ( as suggested by many works
; e.g. , @xcite ) further increases this discrepancy , while small changes of the coefficient in eq .
[ equi : modellx ] are insufficient to compensate it ( e.g. , @xcite adopted a value of 2.5 ) .
the small photometry region of our sample galaxies also seems insufficient to explain this large discrepancy , because the inner region of a halo often dominates the x - ray emission owing to the high density and metallicity of the hot gas @xcite .
we further highlight the oversimplicity of the pure accretion model by comparing the halo gas radiative cooling rate ( @xmath172 ; paper i ) to the sfr inferred from the ir luminosity . as shown in fig .
[ fig : analyticalmodel]b , the @xmath173 relation can be characterized with a linear function : @xmath174 indicating that the radiative cooling of the x - ray emitting corona accounts for only a small fraction of the current sfr .
this is a remarkably different situation from observed in many massive elliptical galaxies , for which the radiative cooling rate is often far greater than the current sfr ( e.g. , @xcite ) .
the above analytical model assumes that only the accretion of gas that has cooled via x - ray emission in the vicinity of galactic discs replenishes the gas for the disc sf . in reality , other fueling mechanisms , such as the accretion without cooling via soft x - ray emission , or the recycling of gas from existing stellar populations , are likely efficient to replenish the cool gas consumed in sf . in particular , @xcite found that the recycled gas from stellar mass loss can provide most or all of the fuel required to sustain the current level of sf in late - type galaxies .
in addition , gas accretion is thought to be bimodal , with maximum past temperatures either of order the virial temperature or @xmath175 ( e.g. , @xcite ) . in general , the cold mode accretion ( at a low temperature )
is believed to dominate the growth of low mass galaxies with baryonic mass typically @xmath176 @xcite .
furthermore , @xcite have found that for the accretion onto galaxies ( as opposed to accretion onto halos ) , cold mode is always significant and the majority of stars present in any mass halos at any redshifts were formed from gas accreted in cold mode .
the discrepancy between the x - ray radiative cooling rate and the current sfr ( eq .
[ equi : rcoolsfr ] ) apparently indicates a higher than unity sf efficiency @xmath171 , which further indicates the existence of additional gas sources as discussed above .
but even if we consider these gas sources ( i.e. , by substituting eq .
[ equi : rcoolsfr ] into eq .
[ equi : modellx ] ) , eq .
[ equi : modellx ] still tends to over - predict the coronal luminosity by a factor of five .
it is thus obvious that the pure accretion model _ can not _ reproduce the observed x - ray luminosity , as previously pointed out by @xcite and others ( e.g. , @xcite ) .
we next compare the x - ray measurements of our _ chandra _ sample and the massive disc galaxies from the literature ( [ paperiiisubsec : massivegalaxy ] ) to the results from the gimic simulations , which explicitly model both the accretion of gas in a cosmological framework and the feedback that results from the formation and evolution of stars . here
we are careful to compare the coronal luminosity ( @xmath149 ) of the observed and simulated galaxies within the same x - ray bandpass and the same radial aperture size ( [ paperiiisec : data ] ) .
a quantitative and uniform characterization of the thermal and chemical states ( temperature and metal abundance ) of the coronal gas is significantly more complicated .
readers interested in such issues may refer to @xcite or our discussions in papers i and ii . as shown in fig .
[ fig : comparegimic ] , @xmath149 of the gimic galaxies better matches the x - ray observations than the simple accretion model predictions of [ paperiiisubsec : analytical ] .
in fact , gimic reproduces not only the coronal x - ray emission in @xmath134 of @xmath177 galaxies , but the scatter in @xmath149 at a given @xmath11 or @xmath148 as well ( fig .
[ fig : comparegimic]a , c ) .
the substantial reduction of the coronal luminosity in the gimic simulations as compared to the pure accretion model is due to the mass reduction of hot gas within a dark matter halo , as a result of the selective cooling and dropout of low entropy gas ( e.g. , via sf ) and galactic feedback @xcite .
the observed and simulated galaxies generally overlap with each other in the @xmath178 plane ( fig . [ fig : comparegimic]a ) .
the most significant outlier is ngc 266 , for which the inclination correction of the rotation velocity is poorly constrained ( [ paperiiisubsec : massivegalaxy ] ) .
furthermore , starburst galaxies in our _ chandra _ sample ( the red data points ) also appear to be systematically more x - ray luminous ( at a given @xmath11 ) than non - starburst ones , as well as the gimic galaxies .
the dark matter halo dominates the potential on large scales in a galaxy .
@xcite have recently found a tighter correlation of the coronal luminosity with the total mass ( than with the stellar mass ) for a sample of early - type galaxies .
this result is qualitatively consistent with the @xmath178 correlation shown in fig .
[ fig : comparegimic]a , although for different types of galaxies .
this consistency indicates , as is clearly implied by gimic , that the halo mass may be a primary factor in the retention of hot gas in general .
therefore , we further convert the rotation velocity to the dark matter halo mass for each of our observed sample galaxies , using the @xmath179 relation of @xcite ( [ paperiiisec : data ] ) and plot @xmath149 v.s . @xmath2 in fig .
[ fig : comparegimic]b .
this halo mass estimation differs from those using the stellar mass - halo mass relation ( smhmr ) obtained from the abundance matching technique ( e.g. , @xcite ) , especially at the high mass end where the large halo mass in smhmr is mostly for groups / clusters of galaxies ( fig .
[ fig : m200scaling]a ) . since the starburst galaxies in our _ chandra _ sample and the four massive disc galaxies from the literature appear to be significantly more x - ray luminous than the non - starburst galaxies in our sample , we fit only these low - mass non - starburst galaxies to obtain a baseline of the @xmath0 relation , which can be characterized as a linear function : @xmath180 in contrast , the average @xmath6 values of starburst and massive disc galaxies are @xmath181 and @xmath182 , or about 13 and 19 times higher than that predicted by eq .
[ equi : lxm200nonstarburst ] .
the gimic galaxies , however , exhibit a much steeper @xmath0 relation , although they are broadly consistent with the observed @xmath178 relation ( fig .
[ fig : comparegimic]a ) .
the @xmath0 relation of the gimic galaxies not only have a much steeper slope ( @xmath183 ) , but also have a much lower scatter ( @xmath184 ) than those of the observed galaxies ( @xmath185 for the low - mass non - starburst galaxies ) .
the @xmath10 correlation is poor for low - mass non - starburst galaxies , but can still be generally described with a linear function : @xmath186 similarly , the average @xmath5 values of starburst and massive disc galaxies are much higher , i.e. , @xmath187 and @xmath188 , or about 7 and 19 times higher than that predicted by eq .
[ equi : lxmstarnonstarburst ] .
in contrast to the @xmath0 and @xmath10 relations , both the starburst and non - starburst galaxies in our _ chandra _ sample can be well described with an identical @xmath1 relation ( fig .
[ fig : comparegimic]d ) , which can also be characterized with a linear function : @xmath189 the coefficient here is only @xmath190 of that for the @xmath1 relation of the entire corona measured by our _ chandra _ observations ( including the contribution from the inner region ; paper ii ) , but the @xmath1 correlation is still good [ spearman s rank order coefficient ( paper ii ) @xmath191 .
the massive disc galaxies again show a large departure from the fitted relation , although they have sfrs comparable to the low - mass starburst galaxies .
the observed galaxies with low sfrs ( e.g. , @xmath192 ) are typically dwarfs , which are excluded from the gimic sample . besides these dwarf galaxies , the gimic galaxies still show significant differences from the observed sample in the @xmath1 plane , i.e. , the fitted @xmath1 index of the gimic galaxies ( @xmath193 ; for galaxies with @xmath194 ; the green solid line in fig .
[ fig : comparegimic]d ) is clearly larger than that of the observed sample ( linear , eq .
[ equi : lxsfr ] ) .
furthermore , there are also some gimic galaxies with high coronal luminosities ( @xmath195 ) but just moderate sfrs ( @xmath196 ) , which are not seen in our observed galaxy sample .
this is potentially an indication that massive galaxies are not being adequately quenched by feedback in the simulations .
as introduced in [ paperiiisubsec : simulations ] , the gimic simulations do not include agn feedback .
the absence of agn feedback in gimic may result in the over - prediction of the number density of _ massive _ disc galaxies @xcite .
the detailed operation of agn feedback on galaxy - wide scales remains ill - understood , but is expected to be important in galaxies with massive spheroids , where the energy output could heat and eject gas , reducing the emissivity in the central regions ( e.g. , @xcite and references therein ) .
we have adopted a high - mass cut of @xmath197 to the gimic galaxies to minimize the effect of missing agn feedback ( [ paperiiisubsec : simulations ] ) .
generally , this agn effect is less important than type ia sne feedback for galaxies with @xmath198 ( e.g. , @xcite ) , and should not qualitatively affect the comparison between observation and simulation in the mass range of our _ chandra _ sample . for galaxies with @xmath199 , however , as shown in fig .
[ fig : m200scaling]a , the stellar masses at fixed @xmath2 are inconsistent with estimates derived from observational techniques @xcite .
this over - prediction of @xmath148 is most likely a result of the enhanced radiative cooling and sf in these massive galaxies due to the absence of agn feedback in gimic .
correspondingly , this missing agn feedback may also explain the high luminosity of coronal x - ray emission at a given @xmath2 , for @xmath200 ( fig .
[ fig : comparegimic]b ) . to first order , the rotation velocity of galaxies is determined by the mass and concentration of their parent halo . whilst the concentration of halos modeled in collisionless simulations is known to follow a tight relation ( e.g. , @xcite ) , the co - evolution of baryons and dark matter in more realistic simulations induces systematic deviations and scatter in this relation ( often described as a `` back reaction '' , e.g. , @xcite ) . as shown in fig .
[ fig : m200scaling]b , low - mass galaxies ( @xmath201 ) from gimic exhibit a larger scatter in @xmath11 than more massive counterparts .
the low-@xmath11 boundary of low - mass galaxies is roughly consistent with @xcite s relation ( derived from collisionless simulations ) , and is used here to estimate @xmath2 of all the observed galaxies except for ngc 266 . on the other hand , the high-@xmath11 boundary of low - mass galaxies seems consistent with high - mass ones .
in contrast , high - mass galaxies with @xmath202 have clearly lower scatter in @xmath11 , and no galaxies have @xmath11 as low as predicted by @xcite s relation . in the absence of agn feedback
, the halo of a massive galaxy will tend to over - cool and over - contract , resulting in a greater rotation velocity .
therefore , the absence of agn feedback in gimic , particularly for high - mass galaxies , likely results in an artificially high baryonic matter density at small radii , and thus a higher dark matter density and maximum rotation velocity . our sample is selected based on the available _ chandra _ data .
archival observations are often proposed for active starburst galaxies which tend to have luminous x - ray halos . among the 30 galaxies used in the present paper , @xmath203
are classified as `` starburst '' ( @xmath204 if one includes the four massive disc galaxies ; table [ table : galaxypara ] ) , while in the gimic _ sample _ ( not necessarily the simulation taken as a whole ) , this fraction is only @xmath205 ( defining a simulated galaxy with @xmath206 as a starburst , broadly consistent with the definition for the observed galaxies ; fig .
[ fig : sfproperties]a ) .
this low fraction in the sample is likely a consequence of the exclusion from the sample of interacting galaxies .
we thus separated starburst and non - starburst galaxies in the above comparisons ( [ paperiiisubsec : comparegimic ] ) .
the coronal luminosity discrepancy between the observed starburst and the gimic galaxies may be partly attributed to this difference in the sample selection . however , at small radii , the x - ray intensity always drops for gimic galaxies @xcite , but increases for the observed ones ( paper i ) . as presented in [ paperiiisubsec : comparegimic ] , the coronal luminosity of our _ chandra _ sample galaxies in @xmath134 is on average @xmath190 of those adopted in papers i and ii ( [ paperiiisubsec : comparegimic ] ) , while the corrected coronal luminosity of the gimic galaxies in @xmath134 is typically @xmath163 of the luminosity in the entire halo ( [ paperiiisubsec : simulations ] ) .
considering that the x - ray luminosity adopted in papers i and ii is also measured in the inner region , we could conclude that gimic under - predict the coronal x - ray emission of low - mass galaxies at least at small radii , although the chosen of photometry region of @xmath161 has minimized this discrepancy .
in fact , this under - prediction may also exist at @xmath134 , as indicated by the steeper slope of the @xmath1 relation for the gimic galaxies ( with respect to the observed galaxies in our _ chandra _ sample ; fig .
[ fig : comparegimic]d ) .
we consider the sfr range of @xmath207 , because galaxies with larger sfrs are likely massive ones in gimic and starburst in our _ chandra _ sample , while those with a lower sfr in gimic have too large scatter in @xmath149 and many of them are indeed massive galaxies ( see below ; fig .
[ fig : comparegimic]d ) .
we find an average coronal luminosity of @xmath208 for our observed galaxies , compared to @xmath209 for the simulated ones . the most intuitive reason for this under - prediction of x - ray luminosity around @xmath177 non - starburst field galaxies may be from the adopted physical and/or numerical ingredients in gimic .
many physical assumptions and parameters , as well as the adopted numerical methods ( e.g. , @xcite ) , may affect the simulated galaxy properties , e.g. , the adopted feedback implementation , metal line cooling , stellar initial mass function , and the assumed cosmology ( e.g. , @xcite ) . in particular , the choice of feedback parameters plays a key role in reproducing the observed sf properties of galaxies ( e.g. , @xcite ) , and possibly the related halo gas cooling and x - ray emission . as suggested in many cosmological simulations , there exists a fundamental problem that the specific sfrs of galaxies ( particularly low - mass galaxies ) at low - redshift are significantly under - predicted ( e.g. , @xcite ) .
it has been suggested that this discrepancy arises because both semi - analytic and hydrodynamical simulations couple the growth of galaxies too strongly to the growth of their host dark matter haloes .
since haloes assemble early , simulated galaxies do so too , and thus form significantly earlier than is observed in nature ( e.g. , @xcite ) .
similarly , gimic also under - predicts the sfr of low - mass galaxies @xcite .
this under - prediction of current sf activity may have potential effects on the x - ray emission , e.g. , via related stellar feedback and cool - hot gas interaction , but the net effect is not yet well understood .
for example , there exists a small number of low - sfr high-@xmath149 galaxies in the gimic simulations ( fig . [
fig : comparegimic]d ) . as indicated in fig .
[ fig : sfproperties]b , c , these galaxies are most likely massive ones , which are apparently post - starbursts whose sf have already ceased but the radiative cooling of the coronal gas is still strong .
some observed galaxies ( e.g. , ngc 2787 , ngc 2841 , ngc 4594 , and ngc 5170 ) may have similar sf properties , i.e. , with low sfr but high @xmath148 .
the sf activities in these galaxies have probably been quenched , e.g. , via morphological quenching or other mechanisms ( e.g. , @xcite and references therein ) . however , the diffuse x - ray luminosities of these galaxies are much lower than those of other massive galaxies with higher sfr ( e.g. , the massive disc galaxies from the literature ; fig . [
fig : comparegimic ] ) , apparently inconsistent with what is predicted by gimic .
the above comparisons clearly demonstrate that x - ray observations can place important constraints on the galaxy formation theory , especially its feedback implementation .
improved feedback prescriptions , especially in low - mass galaxies , may need to be considered in next generation simulations to correctly reproduce both the current sfr and the coronal x - ray luminosity .
in particular , gimic adopted constant kinetic feedback parameters ( initial wind velocity of @xmath210 and mass loading factor of @xmath211 ) , which were chosen to scale the global sfr density to correspond to observational estimates .
although the adopted efficiency ( the wind energy accounts for @xmath212 of the sn energy ) and mass loading factor are roughly in agreement with observational estimates [ e.g. , for m82 by @xcite ] , it is likely that the macroscopic properties of outflows correlate with the detailed structure of the ism ( e.g. , @xcite ) and , by extension , the properties of galaxies .
the simplicity of the adopted physical and/or numerical ingredients in gimic is a possible reason that the sf and x - ray properties of the simulated galaxies are too strongly coupled with the dark matter halo mass of the galaxy ( e.g. , fig . [
fig : comparegimic]b )
. it will be interesting to compare the x - ray scaling relations of gimic galaxies with those of galaxies formed in next - generation hydrodynamical simulations adopting feedback schemes calibrated to reproduce the stellar mass function of local galaxies .
another potential explanation for the under - prediction of soft x - ray emission is the adoption of a single - phase ism in gimic .
the gas is assigned an average temperature within the numerical resolution , which is typically insufficient to resolve a single massive sf region . in these sf regions ,
cold gas often dominates the total gas mass , while hot gas dominates the x - ray emission .
therefore , averaging the thermal state of the gases will result in a too low temperature to emit x - ray efficiently , thus significantly under - predict the x - ray emission .
this effect is the most significant near a galactic disc , where the cool / hot gases strongly mix with each other , and the x - ray emission can be enhanced via enhanced radiative cooling of mass loaded cool gas and charge exchange ( e.g. , @xcite ) .
this effect may also help to explain the declining x - ray intensity profile toward small radii in gimic @xcite , which is not typically observed in real galaxies .
as a general conclusion , gimic likely under - predict the role of sf in x - ray emission for low mass galaxies .
there are two origins of this under - prediction : the under - prediction of the sfr for low - mass galaxies due to the too strong coupling of the growth of galaxies to the growth of their host dark matter halos , and the under - prediction of x - ray emission from sf due to the assumption of the single - phase ism .
although the @xmath0 or @xmath10 slopes themselves are not well constrained due to the large scatter of the observational data , it is clear that the four massive disc galaxies ( ngc 266 , ngc 1961 , ngc 6753 , ugc 12591 ) are more x - ray luminous than the non - starburst galaxies below a typical transition mass of @xmath3 or @xmath4 ( fig . [
fig : comparegimic]b , c ) . as
argued in previous studies ( e.g. , @xcite ) , this high x - ray luminosity is most likely a signature of the accretion of intergalactic gas .
there are several possible mechanisms which may produce the high x - ray luminosity of these massive disc galaxies .
_ firstly _ , the deeper gravitational potential , and possibly the stronger surrounding thermal and ram - pressure ( e.g. , @xcite ) , may help the galaxy to retain more hot gas within its halo , resulting in an increasing baryon fraction ( especially the hot gas fraction ) with the halo mass ( e.g. , @xcite ) . _
secondly _ , the change of the @xmath5 ratio could also be a result of the change of the accretion mode .
numerical simulations have confirmed the change of the cold to hot accretion from below to above the ( baryonic ) mass of @xmath213 of a galaxy ( e.g. , @xcite ) . above this transition mass ,
a larger fraction of gas can be gravitationally heated to an x - ray emitting temperature in the galaxy vicinity .
_ finally _ , the dynamical state of the coronal gas may also affect its overall x - ray emissivity by changing its radial distribution @xcite . a massive galaxy with a corona that tends to stay in a hydrostatic or even inflow
state can have a steeper density profile ( higher gas density in the center ) , enhancing the x - ray emission .
conversely , a galaxy with its corona in a subsonic outflow state tends to have a flatter density profile and may be less luminous in the galactic inner regions ( e.g. , @xcite ) . in all cases , the dark matter halo mass appears to be a key parameter in determining the coronal luminosity and it is likely that the accretion of intergalactic gas plays an increasingly important role in producing the galactic coronae above the transition mass .
in order to study the coronae around disc galaxies , we have constructed a _
chandra _ database of 53 nearby highly - inclined disc galaxies ( paper i ) , and have conducted a correlation analysis of their coronal and other multi - wavelength properties ( paper ii ) . in this paper , we have compared our results to the predictions from a simple analytical model considering only the accretion of intergalactic gas ( in the absence of feedback ) , the measurements of several massive disc galaxies from the literature , as well as results from recent cosmological hydrodynamical simulations invoking both accretion and feedback ( the gimic simulations ) .
our main results and conclusions are set out below .
the observed x - ray emission in the vicinity of galactic discs ( those explored in our _ chandra _ observations without further corrections in [ paperiiisubsec : oursample ] ) is much less ( @xmath170 ) than predicted by a simple analytical model considering only the accretion of intergalactic gas .
furthermore , the radiative cooling rate of coronal gas in the same radial aperture is typically too low ( @xmath214 of the sfr ) to be the primary source of gas to maintain the on - going sfr in the galaxies .
the gimic simulations more accurately reproduce the observed x - ray scaling relations .
they broadly reproduce the luminosity range of the coronal x - ray emission in @xmath134 of @xmath177 galaxies , including the scatter in @xmath149 at a given @xmath11 or @xmath148 .
however , the @xmath0 relation of simulated galaxies differs from that inferred from observations , both in terms of the slope and scatter . for the observed galaxies , low - mass starbursts appear to be more x - ray luminous than more quiescent counterparts at a given galaxy mass .
there is a tight @xmath1 correlation , even for corona at large radii of @xmath161 , similar to that of the x - ray emission at smaller radii .
the overall trends of the observed and simulated galaxies on the @xmath1 plot show little similarity , indicating that gimic potentially under - predicts the role of sf in producing the x - ray emission .
there are two possible origins of this under - prediction : the under - prediction of the sfr for low - mass galaxies due to the too strong coupling of the growth of galaxies to the growth of their host dark matter halos , and the under - prediction of x - ray emission from sf due to the assumption of the single - phase ism .
coronal x - ray luminosity increases much faster with galaxy mass above a typical transition mass of @xmath3 or @xmath4 . below this mass ,
the @xmath0 and @xmath10 relations of the observed galaxies can both be well characterized with a linear scaling relation .
the higher @xmath5 or @xmath6 ratio of galaxies above the transition mass indicates that a massive disc galaxy tends to have a massive dark matter halo , to be dominated by hot - mode accretion , and/or to host a corona most likely in a hydrostatic state .
the accretion of intergalactic gas likely plays an increasingly important role in producing the galactic coronae with increasing galaxy mass .
the above results demonstrate that x - ray observations can place important constraints on the galaxy formation theory , especially the astrophysics of accretion and feedback .
although the simulations have made a great progress to reproduce the observed x - ray properties of galaxies , some significant discrepancies still exist and are quite suggestive .
the absence of agn feedback in gimic tends to over - predict the number of massive galaxies , as well as a simultaneous over - prediction of @xmath11 , @xmath148 , and @xmath149 at a given @xmath2 for massive galaxies .
in addition , gimic may yield too strong a coupling of the galaxy and coronal properties ( @xmath11 , @xmath148 , sfr , @xmath149 ) to the host dark matter halo mass , which is one of the major origins of the discrepancies between the observations and simulations .
furthermore , the adoption of constant feedback parameters ( mass loading and injection velocity ) is likely too simplistic .
it is also important to study the x - ray properties of low - mass galaxies , which are not massive enough to retain hot coronae , and the x - ray emission is mostly from stellar feedback and the related cool - hot gas interaction , which is poorly understood from current numerical simulations .
the present paper is based on the archival observations and x - ray measurements from the literature .
most of these observations are too shallow to probe the faint x - ray emission from the outskirts of the galactic halos .
in addition , the sample is still not large nor uniform . in particular
, very few galaxies are observed in x - rays around or above the transition mass .
we have also focused on the _ x - ray luminosities _ of the galactic coronae .
most of the archival observations are too shallow to yield valuable constraints on hot gas abundance , and hence on gas density and total baryon content of the galaxies ( e.g. , @xcite ) .
therefore , deeper x - ray observations of more galaxies spanning the entire galaxy mass range are highly desirable in order to put tighter constraints on the galaxy formation models and further improve our understanding of the global properties of galactic coronae .
the authors thank @xmath215kos bogd@xmath153n for helping us to compute the x - ray luminosity of ngc 266 in @xmath134 , ying zu and @xmath215kos bogd@xmath153n for many helpful discussions , the anonymous referee for many useful suggestions that lead to improvements of the paper , as well as the referees of papers i and ii whose constructive comments and suggestions led to the write - up of the present paper .
jiang - tao li acknowledges the financial support from cnes and the support from nsfc grant 11233001 . | x - ray observations of circumgalactic coronae provide a valuable means by which to test galaxy formation theories .
two primary mechanisms are thought to be responsible for the establishment of such coronae : accretion of intergalactic gas and/or galactic feedback . in this paper
, we first compare our _ chandra _ sample of galactic coronae of 53 nearby highly - inclined disc galaxies to an analytical model considering only the accretion of intergalactic gas .
we confirm the existing conclusion that this pure accretion model substantially over - predicts the coronal emission .
we then select 30 field galaxies from our original sample , and correct their coronal luminosities to uniformly compare them to deep x - ray measurements of several massive disc galaxies from the literature , as well as to a comparable sample of simulated galaxies drawn from the galaxies - intergalactic medium interaction calculation ( gimic ) .
these simulations explicitly model both accretion and supernovae feedback and yield galaxies that exhibit x - ray properties in broad agreement with our observational sample .
however , notable and potentially instructive discrepancies exist between the slope and scatter of the @xmath0 and @xmath1 relations , highlighting some known shortcomings of gimic , for example , the absence of agn feedback , and possibly the adoption of constant stellar feedback parameters .
the simulated galaxies exhibit a tight correlation ( with little scatter ) between coronal luminosity and halo mass .
having inferred @xmath2 for our observational sample via the tully - fisher relation , we find a weaker and more scattered correlation . in the simulated and observed samples alike , massive non - starburst galaxies above a typical transition mass of @xmath3 or @xmath4 tend to have higher @xmath5 and @xmath6 than low - mass counterparts , indicating that the accretion of intergalactic gas plays an increasingly important role in establishing the observable hot circumgalactic medium with increasing galaxy mass . |
we consider a wiretap set - up , in which a message is transmitted to its legitimate receiver bob in the presence of eve the eavesdropper .
eve is assumed to have unlimited computational power , but to experience an additional noise compared to bob .
lattice coset coding is utilized to maximize eve s confusion , cf .
bob s lattice is referred to as the code lattice or dense lattice , and eve s lattice as the sparse or coarse lattice .
the channel is assumed to exhibit additive white gaussian noise ( awgn ) but no fading .
the respective channel equations for bob and eve are @xmath0 where @xmath1 is the received vector , @xmath2 the transmitted coset - coded vector , and @xmath3 is awgn with respective variances @xmath4 . in finding optimal lattice wiretap codes , there are three main objectives : * maximizing the data rate @xmath5 , which is determined by the size of the codebook @xmath6 and the decoding delay @xmath7 as @xmath8 bits per channel use ( bpcu ) . * minimizing the legitimate receiver s decoding error probability . * minimizing the eavesdropper s probability of correct decision . considering only the first two problems , the largest codebooks for a fixed transmission power and an upper bound for the receiver s error probability are the solutions to the widely investigated sphere - packing problem .
this results in lattices that are typically nonorthogonal ( see , e.g. , @xcite ) .
orthogonal lattices have still traditionally been preferred due to an easy - to - implement bit - labeling algorithm , namely the gray - mapping . due to this mapping ,
the encoding and decoding procedures are more straightforward for orthogonal lattices than for nonorthogonal , i.e. , skewed lattices .
nevertheless , computationally efficient closest - point algorithms such as the sphere decoder also exist for nonorthogonal lattices ( for an explicit construction , see @xcite , sec . 4 ) . in @xcite
, it was also demonstrated how skewed lattices can be efficiently encoded and decoded by using a modified power - controlled sphere decoder or sphere decoding adjoined with minimum - mean - square - error generalized - decision - feedback - equalization ( mmse - gdfe ) , both resulting in optimal ( maximum - likelihood ) performance . hence , skewed lattices should not be excluded when searching for optimal lattices , in particular in the light of the present paper showing that they are not only better in terms of bob s performance , but also in terms of confusing the eavesdropper .
similarly to the sphere - packing problem in bob s case , we now include the third objective in our consideration .
our approach is to fix the data rate and the transmission power and then compare skewed and orthogonal lattices from the point of view of the latter two objectives in an awgn channel .
we study an expression that has two alternative interpretations as either the receiver s error probability ( rep ) for any lattice code in an awgn channel or the eavesdropper s correct decision probability ( ecdp ) for a lattice coset code in an awgn channel .
we prove the following results ( notation will be defined in the subsequent section ) . * skewing bob s orthogonal code lattice @xmath9 will decrease the rep of any code . * skewing eve s orthogonal sparse lattice @xmath10 will decrease the ecdp of any lattice coset code . * combining the previous two results , the common set - up of the dense lattice @xmath9 being orthogonal and the commonly used choice of an orthogonal sublattice @xmath11 are suboptimal in terms of both the ecdp and the rep . according to whether gray - labeling is insisted or not , this common set - up can be improved by either choosing a skewed sublattice of the same orthogonal dense lattice @xmath9 , leaving bob s lattice orthogonal and the rep suboptimal , or skewing both lattices .
these results suggest that skewed lattices deserve more attention in the study of the awgn wiretap channels even though their encoding and decoding are admittedly somewhat more complicated than that of orthogonal lattices .
it is also worthwhile to keep in mind that in any practical system , an outer error correcting code , e.g. , a low - density parity - check ( ldpc ) code , is used in addition to the inner lattice code .
the true decoding bottle - neck in this case is the outer code requiring soft input , not the lattice code .
in this section , we present some necessary definitions and their information - theoretic interpretations .
a _ lattice _ is a discrete additive subgroup of @xmath12 .
any point in a lattice @xmath13 can be expressed in terms of a _ generator matrix _ @xmath14
as follows @xmath15 we assume that the columns of @xmath16 are linearly independent over @xmath17 and hence , the _ lattice coordinates _
@xmath18 of a lattice point are unique .
if @xmath19 , the lattice is of _ full rank_. a _ sublattice _ of a lattice of dimension @xmath20 in @xmath12 is an additive subgroup ; it has a generator matrix @xmath21 , where @xmath22 . here
@xmath23 is the dimension of the sublattice and for a square matrix @xmath24 , @xmath25 the _ volume _ @xmath26 of the lattice @xmath27 is the volume of the fundamental parallellotope spanned by the column vectors of @xmath16 , given by @xmath28 for full - rank lattices .
differing from some information theory references , here vectors are identified with _ column _ matrices and the lattice generator vectors with the _ columns _ of the generator matrix @xmath16 . the _ dual lattice _
@xmath29 of a full - rank lattice @xmath27 generated by @xmath16 is the one generated by @xmath30 let @xmath27 be a full - rank lattice with generator @xmath16 and let @xmath31 be a continuous function with @xmath32 and @xmath33 such that the partial sums of @xmath34 converge uniformly whenever @xmath35 is restricted onto a compact set .
then , @xmath36 where the fourier transform is defined as @xmath37 the proof is given in @xcite .
we point out that the condition on the continuity of @xmath38 is essential for the proof and is missing in the book .
the function that we will optimize is the following .
the _ psi function _
@xmath39 of a lattice @xmath27 at a point @xmath40 is given by @xmath41 this is a variant of lattice theta series restricted on the imaginary axis , @xmath42 .
the convergence properties of the psi series follow from those of the theta series . in @xcite , an upper approximation for the ecdp @xmath43 for a lattice coset code
is derived as @xmath44 here @xmath9 is the dense and @xmath10 the sparse lattice , inteded for the receiver and the eavesdropper , respectively .
the lattices are assumed to be of full rank and the eavesdropper s noise is assumed to be awgn with variance @xmath45 .
the inequality is tight for large @xmath46 . for small @xmath46 ,
the upper bound is larger than @xmath47 and hence useless . on the other hand ,
using the union bound technique as is done in ( * ? ? ?
* appendix ii ) for rayleigh - fading channels and setting the rayleigh fading coefficients equal to one , the rep can be approximated from above as : @xmath48 this formula is valid for any lattice code @xmath9 ( not just a coset code ) in an awgn channel and the approximation is good for small receiver s noise variances @xmath49 . based on these two formulae and the fact that the variances @xmath45 and @xmath49 vary with the random channels , our subsequent aim will be to provide inequalities of the form @xmath50 for all @xmath40 . when comparing different lattices sharing the same dimension , their volumes are first normalized to one .
this ensures that for a relatively large fixed transmission power , the finite codebooks carved from the infinite lattices will be approximately equally large , and hence we can fairly compare the lattice codes without considering the actual data rates , as these will coincide . due to the obvious connection between the formulae and for the rep and ecdp ,
respectively , one would intuitively guess that a solution for the sphere - packing problem also yields an optimal ecdp .
this , however , does not seem to work on the level of mathematical proofs ; eq .
is obtained by the union bound technique , whereas in the sphere - packing problem , the upper bound for rep is based on integrating a gaussian function over a ball , yielding a much tighter bound for large receiver s noise variances @xmath51 or , equivalently , for small arguments of @xmath52 . to minimize the ecdp , we want to minimize @xmath52 for small arguments . hence ,
even if the sphere - packing probability bound is small , it does not provide us with immediate information as to how small the @xmath52 function is for small arguments , i.e. , how small the ecdp is .
in this section , we show that skewing a lattice will always improve a code both in terms of eve s and bob s probabilities .
[ translation lemma ] for any full - rank lattice @xmath27 , @xmath53 for any @xmath54 .
denote summands of the respective sides as @xmath55 and @xmath56 , so @xmath57 .
then , by the elementary properties of fourier transform , we have @xmath58 .
hence , using the poisson formula , @xmath59 where @xmath16 is the generator matrix of @xmath27 . in continuation , we will use the knowledge that the fourier transform of the gaussian function @xmath60 is another gaussian , hence a real , positive and even function .
( the explicit form of @xmath61 could be calculated but it is not necessary . ) first , since @xmath61 is even , the imaginary parts @xmath62 of the summand for lattice @xmath29 points @xmath63 cancel out , yielding @xmath64 next , we need the positivity of @xmath61 to be able to approximate the cosine by @xmath47 . first , note that we assumed @xmath54 , equivalently , @xmath65 with some component of @xmath66 , say the @xmath67 one @xmath68 , not integer . also note that @xmath69 where @xmath70 . hence , @xmath71 .
now , choosing the lattice point @xmath72 such that @xmath73 , we immediately see that @xmath74 and @xmath75 .
hence , replacing @xmath76 by @xmath47 in the preceding step , we get a strict inequality @xmath77 where we have again applied the the poisson formula to .
finally , the double fourier transform is in general a reflection operator , so @xmath78 , and using the fact that @xmath79 we obtain the result , @xmath80 [ def : skew ] let @xmath81 be a full - rank orthogonal lattice in @xmath12 with generator vectors @xmath82 , @xmath83 for @xmath84 .
we call a lattice @xmath85 a _ skewing _ of @xmath81 , if it has a generator matrix that is an upper triangular matrix with the diagonal elements @xmath86
. this definition has a simple geometric interpretation , depicted in fig . [ skewfig ] .
we point out that skewing can be interpreted as a matrix operation .
if @xmath87 and @xmath88 are the generator matrices of @xmath81 and @xmath89 , respectively , then the non - singular skewing matrix @xmath90 can be solved from the matrix equation @xmath91 this equation is non - singular , since @xmath92 and @xmath93 by the determinant rule of upper triangular matrices .
this also implies that @xmath94 .
now we are ready to state the main theorem .
after this , we will provide an illustrative interpretation of the theorem and prove it . [ skewing theorem ] for a skewing @xmath89 of a full - rank orthogonal lattice @xmath81 , @xmath95 for all @xmath96 .
skewings provide several easy ways to improve lattice coset codes .
we point out that since skewing keeps the lattice volume constant ( @xmath94 in matrix representation ) , it will not affect the size of a spherical codebook .
hence , a lattice comparison between skewings only requires considering the ecdp and the rep . with ths knowledge , the theorem has the following immediate implications . * comparing a dense lattice @xmath97 and its skewings @xmath98 , theorem [ skewing theorem ] applied to eq .
shows that the rep is always smaller for the skewings @xmath98 .
this holds for all codes , not just coset codes .
* consider a coset code arising from a fixed nonorthogonal lattice @xmath9 .
then , to minimize the ecdp , it seems that @xmath10 should not be chosen orthogonal ( if orthogonal sublattices exist ) .
note that then no skewing of the orthogonal @xmath10 is necessarily a sublattice of @xmath9 , so this is just heuristics . * consider a typical set - up of @xmath97 generated by @xmath99 being orthogonal and @xmath100 . in this case both the rep and the ocdp are suboptimal .
there are two remedies : * * first , we can skew both @xmath101 and @xmath97 .
skewing by @xmath90 so that the skewed lattices @xmath102 and @xmath103 are generated by @xmath104 and @xmath105 , respectively , will yield a nonorthogonal lattices but preserve the volumes : @xmath106 ( since @xmath94 ) .
hence , applying this and theorem [ skewing theorem ] in eqs . and , we see that skewing will decrease both the rep and the ecdp .
however , the skewed lattice will not allow for a simple gray mapping , or in other words , the gray mapping is not guaranteed to give an optimal bit - labeling . * * second , we can only opt for skewing the sublattice @xmath10 , while leaving @xmath9 orthogonal .
this means that the rep will remain suboptimal , but the lattice will allow for gray labeling and maintains simpler encoding and decoding for bob along the lines discussed in the introduction .
moreover , the ecdp is decreased .
the idea is that if @xmath97 is orthogonal and generated by @xmath107 , and @xmath108 , then any sublattice @xmath103 generated by @xmath21 , where @xmath24 is an upper triangular integer matrix with diagonal etries @xmath109 , is easily proven to be a skewing of @xmath110 ( or equal to @xmath110 ) .
then , applying theorem [ skewing theorem ] to eq . , we see that @xmath111 will yield a lower ecdp .
let us use the notation of definition [ def : skew ] .
furthermore , denote by @xmath112 the embedding into @xmath113 , @xmath114 , of @xmath115 .
equivalently , @xmath112 is the lattice in @xmath113 generated by the @xmath23 first columns of the generator @xmath88 of @xmath89 . continuing to ease the notation ,
denote the projection of the @xmath116 column of @xmath88 onto @xmath117 by @xmath118 , so the @xmath116 column is @xmath119 .
now , it is apparent from the definition of @xmath120 that @xmath121 and that @xmath122 on the other hand , using lemma [ translation lemma ] for the lattice @xmath123 , @xmath124 , which is a full - rank lattice of @xmath117 , we obtain @xmath125 and , as stated in lemma [ translation lemma ] , the equality holds if and only if @xmath126 for all @xmath127 , equivalently , @xmath128 .
this is furthermore equivalent to that the @xmath116 column @xmath119 of @xmath88 can be replaced by @xmath129 without changing the lattice @xmath89 .
next , starting from eq .
, using the identity inductively , and finally using eq . , we obtain @xmath130 the equality holds if and only if it has been possible to modify , for all @xmath23 , the @xmath116 column of @xmath88 into @xmath131 without changing the lattice generated by @xmath88 . but
this is equivalent to @xmath88 and @xmath132 generating the same orthogonal lattice @xmath81 .
this is impossible by the definition of a skewing .
hence , for any skewing @xmath89 of @xmath81 , we have a strict inequality @xmath95 for all @xmath96 .
this completes the proof . the gosset lattice @xmath133 has the generator matrix @xmath88 given by @xcite @xmath134 so it is a skewing of the orthogonal lattice @xmath27 generated by @xmath135 .
the theta series of the gosset lattice is expressible by the jacobi theta functions as @xcite @xmath136 where @xmath137 .
the theta series of the orthogonal lattice @xmath27 is @xmath138 where @xmath139 and @xmath140 is the @xmath67 diagonal element of @xmath87 .
now , recalling that @xmath42 , we can compare the psi series of these two lattices by evaluating jacobi theta functions .
the plots of the psi functions are depicted in fig .
[ psi_skewings ] .
the figure shows that @xmath141 for all @xmath142 , as predicted by theorem [ skewing theorem ] . in coset coding
, this has the following interpretation : @xmath133 and @xmath27 are both index @xmath143 subgroups of @xmath144 .
if bob s lattice is @xmath144 , then the coset lattices @xmath133 and @xmath27 will yield the same code rates , but @xmath133 with a better secrecy . and its skewing @xmath133.,scaledwidth=50.0% ]
in the construction of lattice codes for awgn wiretap channels , skewed lattices should be taken more seriously .
namely , we have proved that orthogonal lattices are suboptimal not only in terms of the receiver s error probability as we already know from the sphere - packing theorems , but also in terms of the eavesdropper s correct decision probability when using lattice coset codes .
hence , the design of secure lattice codes should ideally be based on skewed lattices .
however , due to implementation purposes , one may opt for only skewing the eavesdropper s lattice , while preserving the orthogonality of the legitimate receiver s lattice , which results in suboptimal performance but easy - to - implement algorithms for bob , as well as improved security .
this work was carried out during a. karrila s msc thesis project .
the department of mathematics and systems analysis at aalto university is gratefully acknowledged for the financial support .
e. viterbo and f. oggier , `` algebraic number theory and code design for rayleigh fading channels '' , _ foundations and trends in communications and information theory _ , vol . 1 , no .
3 , now publishers inc . , 2004 .
j. boutros , e. viterbo , c. rastello and j .- c .
belfiore , `` good lattice constellations for both rayleigh fading and gaussian channels '' , _ ieee transactions on information theory _ , vol .
42 , no 2 , march 1996 . | we consider lattice coset - coded transmissions over a wiretap channel with additive white gaussian noise ( awgn ) . examining a function that can be interpreted as either the legitimate receiver s error probability or the eavesdropper s correct decision probability , we rigorously show that , albeit offering simple bit labeling , orthogonal nested lattices are suboptimal for coset coding in terms of both the legitimate receiver s and the eavesdropper s probabilities . |
there is a growing demand to include partial spatial coherence in optical design , not only because of classical areas of application such as microscopy and projection lithography , but also because of the increasing importance of partially coherent solid - state sources such as multimode lasers ( including excimers ) and leds .
methods based on ray optics can not adequately describe all of the relevant issues related to coherence and polarization , which are intimately connected in electromagnetic coherence theory .
thus wave - optical methods are required , which must be capable of dealing with non - paraxial fields and systems that may contain also micro- and nanostructures in objects and interfaces . to this end
, one needs computationally efficient physical - optics - based representations of spatially partially coherent electromagnetic fields .
apart from some specific models that allow analytic solutions , propagating spatially partially coherent light even in free space is a formidable numerical problem involving four - dimensional integrals @xcite .
the dimensionality of the propagation integrals can be decreased to two ( for planar sources ) if the partially coherent field is represented as an incoherent superposition of fully spatially coherent fields .
the classical way to do this is the coherent - mode expansion of the cross - spectral density ( csd ) function by means of mercer s expansion @xcite , which has recently been extended to electromagnetic fields @xcite . in this representation
the coherent modes are uniquely defined by the csd through a fredholm integral equation ; they form a complete and orthonormal set , in which the effective number of modes @xmath3 increases as the degree of coherence of the field is reduced @xcite . thus , in free - space propagation , the original four - dimensional integral for csd is replaced by @xmath3 two - dimensional integrals for the coherent modes . in light - matter interaction analysis one solves the diffraction or scattering problem for @xmath3 coherent fields of different functional forms @xcite .
there is also an alternative representation of a partially coherent field in terms of uncorrelated , fully coherent fields @xcite . here
all coherent fields or ` elementary modes ' are of identical functional form but spatially ( or angularly ) shifted with respect to each other and weighted by a function determined by the csd . unlike the mercer expansion , the shifted - elementary - mode representation is applicable only to a specific class of genuine csds @xcite .
however , this class contains many of the fields of practical significance in optical design , including all quasihomogeneous fields ( leds , excimers , illumination in microscopy and projection lithography , etc . ) .
since the elementary modes are identical , only a single 2d integral needs to be evaluated in free - space propagation problems . in interaction problems one needs to scan the elementary mode across the object and perform a set of independent diffraction calculations for coherent light .
typically the elementary mode has a smooth functional form , at least compared to the higher - order modes in the mercer expansion , and is therefore easy to propagate numerically .
another advantage of the shifted - mode model is that there is no need for numerical solution of the fredholm integral equation : the elementary mode and the associated weight function can be determined , e.g. , from far - zone properties of the field @xcite . in this paper
we generalize the scalar shifted - elementary - mode representation to the vectorial case , which is necessary to adequately model partially spatially coherent , partially polarized sources .
there are two major reasons why such a generalization is necessary : first , non - paraxial fields can not be adequately described by a scalar model and , second , optical components in the system can modify the polarization state of the field .
it turns out that the vectorial nature of the field does not fundamentally complicate the numerical procedure .
two elementary modes are needed to specify the state of polarization , but also in the electromagnetic case the modes and their weight functions can be determined from far - zone properties of the field .
we begin the discussion by briefly reviewing the scalar model in sect .
[ s : scalar ] to establish the notation and to simplify the interpretation of the main results .
the extension to the electromagnetic case is outlined in sect .
[ s : emext ] and the rigorous mathematical formulation is presented sections [ s : far ] and [ s : elementary ] .
the important special case of rotationally symmetric and quasihomogeneous fields is discussed in sections [ s : rotsym ] and [ s : quasihom ] , respectively .
some numerical results are provided in sect .
[ s : cosine ] . in sect .
[ s : led ] we apply the model to a simple led geometry .
finally , issues such as the measurements required to determine the elementary modes and their weight functions are discussed in sect .
[ s : discussion ] .
using the notations of fig . [
f : notations ] , we may write the well - known relationship @xcite between the cross - spectral density function @xmath4 and the angular correlation function @xmath5 as ( we omit the dependence on the angular frequency @xmath6 for brevity thoughout the paper ) @xmath7 \mathrm{d}^2\kappa_1\,\mathrm{d}^2\kappa_2 . \label{angspecscalar}\end{aligned}\ ] ] here the asterisk indicates complex conjugation , @xmath8 with @xmath9 are position vectors , @xmath10 represent wave vectors , and @xmath11 are their transverse projections .
restricting now to the specific class of fields mentioned in the introduction , we assume that the angular correlation function is of the schell - model form @xcite @xmath12 the radiant intensity of a scalar field is defined as @xcite @xmath13 where @xmath2 is the refractive index of the medium , @xmath14 with @xmath15 is a unit direction vector , @xmath16 is its transverse projection , @xmath1 is the angle between @xmath17 and the @xmath18 axis , and @xmath19 is the wave number . using eq .
, we have @xmath20 thus the radiant intensity of a schell - model partially coherent field defined by eq .
is the same as that produced by a coherent field with angular spectrum @xmath21 .
let us introduce two - dimensional fourier - transform relations @xmath22 and @xmath23 where @xmath24 and @xmath25 is the transverse projection of the position vector .
inserting eq . into eq . and using the fourier representation of @xmath26 obtained by inverting eq . , we get the expression @xmath27 for the csd @xcite .
this result applies , in particular , at @xmath28 .
thus @xmath29 is the coherent source - plane field with angular spectrum @xmath21 .
the representation in eq . expresses the partially coherent field as a weighted linear superposition of spatially shifted but identical fully coherent elementary ( scalar ) fields @xmath29 .
as is evident from eq . , the mathematical form of the elementary field mode @xmath29 can be determined from the knowledge of the radiant intensity ( at least apart from a phase factor , which can in fact be employed to model volume sources @xcite ) . in general
, the weight function @xmath30 can be determined from far - field coherence measurements using eqs . and ,
but often there are simpler ways to at least approximate it @xcite .
this is the case , in particular , if the field is quasihomogeneous , i.e. , if the coherence area in the source plane is much smaller than source area ; in this case the weight function comes out of the integral in eq . . when @xmath29 and @xmath30 are known , coherent propagation techniques for @xmath29 and a linear superposition according to eq .
suffice to propagate the entire spatially partially coherent field .
the extension of the scalar shifted - mode model to the electromagnetic case is not trivial .
one might be tempted to use some superpositions of , e.g. , locally linearly polarized modes of the scalar functional form to model partially polarized or unpolarized sources . however , such constructions seem hard to justify mathematically . the approach taken here
is based on far - field information : in the far zone the field behaves as an outgoing spherical wave , and therefore it has a well - defined local polarization state .
we employ this fact to separate the angular correlation tensor , which is the electromagnetic extension of the function @xmath5 , into two orthogonal parts [ see eq . in sect .
[ s : far ] ] .
these represent the electromagnetic elementary modes ( or polarization modes ) in the far field
. then the source - plane modes can be determined by fourier - transform techniques in analogy with the scalar case . as a result of the construction process ( to be described mathematically in the following sections )
, we obtain a vectorial shifted - mode expansion for the spatially partially coherent field everywhere in space [ see eqs . , , and in sect .
[ s : elementary ] ] . instead of propagating one fully coherent mode as in the scalar case
, we now need to propagate two well - defined fully coherent vectorial field modes , and to form the generalized shifted - mode superposition , to govern the propagation of the electromagnetic spatially partially coherent field .
thus the increase in computational complexity , compared to the scalar case , is essentially fourfold .
a statistically stationary random electromagnetic field in the space frequency domain is described by a csd matrix @xmath31 , which may be expressed as @xcite @xmath32 here @xmath33 indicates the transpose , the brackets denote ensemble averaging , and the electric - field realizations @xmath34 are understood as appropriate random linear superpositions of the eigenfunctions of the fredholm integral equation satisfied by @xmath31 @xcite .
the position - dependent spectral density of the field can be written , in analogy with scalar theory of partial coherence @xcite , as @xmath35 , where tr stands for trace .
the relation between the field at the ( secondary ) source plane @xmath36 and the far - field can be found , for example , using the angular spectrum representation of the csd matrix @xcite .
we thus have , at any plane @xmath37 , @xmath38 \mathrm{d}^2\kappa_1\,\mathrm{d}^2\kappa_2 , \label{angspecdef}\end{aligned}\ ] ] where the angular correlation matrix ( acm ) @xmath39 \mathrm{d}^2\rho_1\,\mathrm{d}^2\rho_2 \label{defangspec}\end{aligned}\ ] ] describes the correlations between vectorial plane - wave components . in the far zone @xcite @xmath40}{r_1r_2
} \label{farfield}\end{aligned}\ ] ] and the poynting vector takes the form @xcite @xmath41 where @xmath2 is the refractive index of the material , and @xmath42 and @xmath43 are the vacuum permittivity and permeability , respectively .
furthermore , the radiant intensity is @xmath44\nonumber\\ & = 2n \pi^2 k^2 \cos^2\theta \sqrt{\frac{\epsilon_0}{\mu_0}}\,\mathrm{tr}\,{\boldsymbol{\mathcal{a}}}(k\boldsymbol{\sigma } , k\boldsymbol{\sigma } ) .
\label{radiantintensity}\end{aligned}\ ] ] owing to eqs . and
, also acm has a representation as a correlation matrix : @xmath45 where the components of @xmath46 represent , componentwise , the angular spectra @xcite of the electric - field realizations .
it follows from eq . that @xmath47 is hermitian , satisfying @xmath48 where the dagger denotes the adjoint matrix .
it is also non - negative definite in the sense that @xmath49 where @xmath50 is an arbitrary , sufficiently well - behaved vector function of the same size as the electric - field vector .
the following argument is essential for the conclusions of this paper : the field in the far zone is well known to be a modulated outgoing spherical wave .
thus , in spherical polar coordinates , the angular - spectrum vector is two - dimensional , i.e. , @xmath51 for every @xmath52 .
hence @xmath53 is expressible as a @xmath54 matrix in these coordinates .
since @xmath53 is a hermitian , non - negative definite @xmath54 matrix , it has two non - negative real - valued eigenfunctions . in view of eqs . and
, we have at @xmath55 the decomposition @xmath56 where @xmath57 are the eigenvalues and @xmath58 are the eigenvectors of @xmath47 .
the eigenvectors may be assumed orthonormal , i.e. , @xmath59 in other words , the matrix @xmath47 can be diagonalized .
evaluation of the trace in eq .
yields the explicit form for the radiant intensity : @xmath60,\label{radiantmodes}\ ] ] where @xmath61 .
thus it is proportional to the sum of the eigenvalues of the acm at @xmath62 . the decomposition in eq .
is , in fact , valid regardless of the chosen coordinate system ; the eigenvalues remain invariant in all unitary transformations , including simple coordinate transformations between , e.g. , cartesian and spherical polar coordinate systems . as explicitly expressed in eqs .
, the polarization decomposition of acm is generally direction - dependent : the eigenvalues and/or the eigenvectors depend on @xmath52 .
the physical meaning of eq .
is clear : in each direction in the far - zone , we may decompose the single - point acm into two mutually uncorrelated , orthogonal polarization components . moreover ,
owing to the factorized form @xmath63 of these components , they both represent fully polarized fields .
this can be verified from the well - known formula for the ( space frequency domain ) degree of polarization : @xmath64 ^ 2}\right\}^{1/2}.\end{aligned}\ ] ] in the far zone we have , using eq .
( [ farfield ] ) , @xmath65 ^ 2}\right\}^{1/2}. \label{farzoned}\end{aligned}\ ] ] with the aid of eqs .
( [ decomposition ] ) and ( [ orthogonality ] ) , we then obtain @xmath66 thus @xmath67 for each individual polarization mode , i.e. , if either @xmath68 or @xmath69 .
it is worth stressing that the polarization decomposition presented above is analogous with the well - known decomposition of a partially polarized plane wave into two polarization modes : in particular , eq .
( [ idegpol ] ) is analogous with eq .
( 6.331 ) in ref . @xcite . in our case , however , the direction - dependent eigenvalues are those of a partially polarized and partially coherent field in the far zone . to be able to include partial coherence ( in addition to partial polarization ) in the analysis , we now assume that also the two - point acm can be expressed in the form @xmath70 where @xmath71 have diagonal values @xmath72 . in other words , @xmath53 is assumed to be expressible as a sum of two mutually uncorrelated , but fully coherent and polarized modes even if @xmath73 .
we stress that , while eq .
( [ assumption ] ) does not follow from eq .
( [ decomposition ] ) , it typically holds .
we do not dwell into a detailed discussion of this point here , but note that it is a nontrivial task to find counterexamples .
we proceed to investigate the angular correlation properties of the class of fields described by eq . .
without loss of generality , we may employ spherical polar coordinates in the far field , in which case @xmath71 is a @xmath54 matrix .
let us denote by @xmath74 a unitary matrix whose columns are the eigenvectors @xmath75 of @xmath47 and by @xmath76 a diagonal matrix with elements equal to the eigenvalues @xmath57 of @xmath77 . then , equivalently with eq . , we may write @xmath78 and consequently @xmath79 in view of eq .
( [ assumption ] ) and the associated discussion , @xmath80 represents the angular correlation between two plane - wave components whose polarization states are described by deterministic vectors @xmath81 and @xmath82 .
thus we may write , in analogy with eq .
( [ decomposition ] ) , @xmath83 where @xmath84 are scalar ( angular correlation ) functions .
it is seen by direct calculation that @xmath53 has a representation similar to eq .
( [ singular11 ] ) , @xmath85 where @xmath86 is a diagonal matrix with elements @xmath84 .
thus @xmath75 and @xmath84 can be determined by singular value decomposition of @xmath87 .
it can be shown , e.g. , using eq . that the singular values
@xmath84 satisfy @xmath88 as a result , we may define the normalized angular correlation functions @xmath89^{1/2 } } \label{normg}\end{aligned}\ ] ] satisfying the inequalities @xmath90 .
if the correlations in the far zone are of ( generalized ) schell - model form , i.e. , @xmath91 , it follows from eqs .
( [ decomposition2 ] ) and ( [ normg ] ) that @xmath92 with @xmath93^{1/2}\mathbf{f}_j ( \boldsymbol{\kappa } ) .
\label{hf}\ ] ] this is the electromagnetic extension of eq . .
it follows from eq . and the linearity of eq . that the csd has the decomposition @xmath94 let us define the inverse fourier transforms @xmath95 and @xmath96 of the functions @xmath97 and @xmath98 in analogy with eqs . and .
with a procedure similar to that used in derivation of eq . , we can express the two terms in eq .
( [ fieldsource ] ) in the form @xmath99 in the half - space @xmath100 . in particular , eq .
( [ elementary ] ) is valid at the source plane @xmath28 , where @xmath101 .
this result is the electromagnetic extension of the scalar elementary - mode decomposition in eq . .
let us next examine some general properties of eq .
( [ elementary ] ) . equations and together with @xmath102 lead to @xmath103 because of the exponential factor in the integral , @xmath104 and @xmath105 are generally not orthogonal in a pointwise sense . for many paraxial fields , though , the functions @xmath106 and @xmath107 are globally orthogonal , i.e. , @xmath108 for all @xmath109 and @xmath110 . in this special case
the pointwise orthogonality of @xmath104 and @xmath105 follows from eqs . , , and .
this property holds also for significant classes of non - paraxial fields , as will be demonstrated in sections [ s : rotsym ] and [ s : cosine ] .
however , it is not essential for practical implementation of the propagation algorithm developed in this paper . since the functions @xmath95 represent fully coherent fields , they obey the helmholtz equation @xmath111 furthermore , the csd matrix obeys the divergence equation @xcite @xmath112 where the subscript 1 denotes differentiation with respect to @xmath113 . together with eqs . and
, this implies that @xmath114 in view of eqs . and , the functions @xmath115 behave exactly as the electric field vector in the space frequency domain .
hence the matrix - functions @xmath116 may be called the _ elementary electric - field modes _ of the field . using these modes , we can write eqs . and in a more compact form @xmath117 and express the spectral density of the field as @xmath118 where @xmath119 are the spectral densities of the two elementary electric - field modes . in analogy with scalar theory @xcite , the field in eq .
is understood to consist of a weighted continuum of identical , laterally shifted elementary modes .
the main difference between the scalar and electromagnetic descriptions is that the electromagnetic field consists of two ( uncorrelated ) sets of elementary modes , whose polarization states in the far - zone are orthogonal and whose radiant - intensity distributions are in general different .
owing to the factorized form of @xmath120 in eq . , each elementary mode is completely coherent @xcite in the sense of the space
frequency analog of the degree of coherence for electromagnetic fields put forward in ref .
@xcite : @xmath121\}^{1/2}. \label{scalarmu}\end{aligned}\ ] ] here @xmath122^{1/2 } } \label{boldmu}\end{aligned}\ ] ] is the normalized csd matrix . in general , all @xmath123 matrix elements of @xmath120 are non - zero and spatially varying .
analogously with eqs . and
, we can define the degree of angular coherence @xmath124\}^{1/2 } \label{scalaralpha}\ ] ] where @xmath125^{1/2 } } \label{boldalpha}\ ] ] is normalized angular correlation matrix . to conclude this section we provide a convenient , fully general series representation for the electric - field modes . in the far zone , where the field is transverse to the local propagation direction and thus fluctuates locally in the plane defined by the spherical polar unit vectors @xmath126 and @xmath127
, it is natural to express the basis vectors in the form @xmath128 using spherical polar coordinates @xmath129 for the spatial frequencies and circular cylindrical coordinates @xmath130 for the position vector ( see fig .
[ f : notations ] ) , we then have ( after somewhat lengthy calculations outlined in appendix [ a1 ] ) @xmath131\nonumber\\ & \qquad+ \boldsymbol{\hat{\phi } } \left [ \frac { m}{\rho}\cos\theta f_{j,\theta , m}(\theta ) -{\rm i } f_{j,\psi , m}(\theta ) \frac{\mathrm{d}}{\mathrm{d\rho } } \right]\nonumber\\ & \qquad- \mathbf{\hat{z } } k \sin^2\theta f_{j,\theta , m}(\theta ) \biggr\}j_m(k\rho\sin\theta)\,\mathrm{d}\theta , \label{cylinder}\end{aligned}\ ] ] where @xmath132 are the azimuthal fourier coefficients of @xmath133 and @xmath134 stands for either @xmath1 or @xmath135 .
note that the upper limit of the integral in eq .
is set to @xmath136 .
thus the elementary electric - field modes are taken to contain only propagating waves , i.e. , information that can be gathered from far - zone measurements . as a result ,
the propagation method considered here is not suitable for modeling near - field phenomena ( fields at distances of the order of one wavelength from the plane @xmath36 ) .
let us assume that the csd at @xmath28 is rotationally symmetric about the @xmath18 axis
. then also the acm is rotationally symmetric and the polarization basis vectors @xmath106 and @xmath107 are rotationally invariant , i.e. , their @xmath1 and @xmath135 components @xmath137 and @xmath138 are independent on the azimuthal angle @xmath135 . hence only the zeroth - order fourier coefficients @xmath139 in eq .
are non - zero and it follows from eqs . and that @xmath140\nonumber\\ & \qquad- \sin\theta j_0(k\rho\sin\theta)f_{j,\theta,0}(\theta)\mathbf{\hat{z } } \biggr\ } \,\mathrm{d}\theta.\end{aligned}\ ] ] thus , as expected , the elementary electric - field modes are also rotationally symmetric about the @xmath18 axis .
in particular , if the basis vectors are parallel to @xmath126 and @xmath127 , @xmath141 \nonumber\\ & \quad\times\exp\left({\rm i } kz\cos\theta \right)\,\mathrm{d}\theta,\\ \mathbf{e}_2(\mathbf{r } ) & = \boldsymbol{\hat{\phi}}\frac{{\rm i}k^2}{2\pi}\int_0^{\pi/2}f_{j,\psi,0}(\theta ) \sin\theta j_1(k\rho\sin\theta ) \nonumber\\ & \quad\times\exp\left({\rm i } kz\cos\theta \right)\,\mathrm{d}\theta.\end{aligned}\ ] ] [ generalradialazimuthal ] the elementary modes @xmath142 and @xmath143 are now radially and azimuthally polarized fields , respectively .
hence they are pointwise orthogonal , regardless of whether the field is paraxial or not .
since these modes have no azimuthal phase variation , they possess no phase singularity ( vortex ) at @xmath144 .
thus , even if the axial field vanished @xmath36 ( as turns out to be often the case for some of the field components ) , such a zero does not propagate .
assume next that the variations of the spectral density @xmath145 at the source plane are slow compared to the variations of the degree of coherence @xmath146 .
moreover , let the correlations at @xmath28 be of the schell - model form , i.e. , depend only on @xmath147 .
such a planar source is said to be quasi - homogeneous @xcite .
we may then approximate @xmath148 , where @xmath149 , and write @xmath150 inserting eq . into eq .
and defining @xmath151 yields @xmath152 where @xmath153 and @xmath154 since @xmath155 it follows from eq .
( [ radiantintensity ] ) that the radiant intensity produced by a quasihomogeneous electromagnetic source depends only on the correlation properties of the source field , in complete analogy with the scalar case @xcite .
inserting from eq . into eq . and
recalling that @xmath156 is a wide function compared to @xmath157 , we have @xmath158 using eq . and noting that @xmath157 is real because @xmath159 is non - negative , we obtain @xmath160 ^ 2\right\}^{1/2}}{{\rm tr}\,\tilde{\mu } ( \bar{\boldsymbol{\kappa}})}. \label{alpha-2}\ ] ] this expression can be cast into a more transparent form using the far - zone degree of polarization defined in eq . , which can be written equivalently in the form @xmath161 ^ 2 } { \left[{\rm tr}\,\tilde{\boldsymbol{\mu}}(\boldsymbol{\kappa})\right]^2}-1\right\}^{1/2}.\ ] ] we then have , from eq . , @xmath162^{1/2}. \label{qhffd}\ ] ]
the first fraction in this expression is equal to the scalar degree of angular coherence . the second fraction , however , is a polarization - dependent modulating term that depends on the source - plane correlations .
the assumption that the source is quasihomogeneous simplifies decisively the elementary - mode decomposition of the field in the scalar case @xcite , and the same is true in the electromagnetic case .
it follows from eqs .
, , and that @xmath163 where , in the last step , we have used eq . .
comparing eqs .
and yields @xmath164 thus the weight function no longer appears inside the integral and , irrespective of the spatial distribution of the spectral density at the source plane , the field characteristics can be determined the propagating a convolution integral involving the elementary modes only .
let us consider the rotationally symmetric case with radially and azimuthally polarized basis vectors @xmath165 , @xmath166 , and corresponding eigenvalues @xmath167 , @xmath168 , where @xmath169 and @xmath170 are arbitrary ( real ) functions of frequency .
then , in view of eq . , the radiant intensity is a superposition of two @xmath0 type contributions , one radially and the other azimuthally polarized : @xmath171.\ ] ] the elementary electric - field modes in the far zone are , according to eq . , @xmath172 using eqs . and
we see that the non - vanishing fourier coefficients are @xmath173 and @xmath174 . inserting these into eqs . and
applying derived in appendix [ a2 ] , we obtain the source - plane elementary field modes in the form @xmath175,\\ \mathbf{e}_2(\rho,0 ) & = \boldsymbol{\hat{\phi}}a_2\frac{{\rm i}k^3\rho}{16\sqrt{\pi } } \gamma\left(\frac{b}{4}\right)\nonumber\\ & \quad\times \sideset{_1}{_2}\operatorname{\widetilde{\mathit{f}}}\left(\frac{3}{2};2,\frac{3}{2}+\frac{b}{4};-\frac{k^2\rho^2}{4}\right),\end{aligned}\ ] ] [ spcosmodes ] where @xmath176 is the gamma function and @xmath177 is the regularized hypergeometric function ( see appendix [ a2 ] ) . figures [ cos1][cos3 ] illustrate the radial dependence of the elementary - field components and the function @xmath178^{1/2}\ ] ] for different values of @xmath179 , with @xmath180 . , as well as the function @xmath181 ( thick solid line ) for @xmath182 , which corresponds to a lambertian source . ] , but with @xmath183 , which corresponds to an incoherent source in scalar theory . ] , but with @xmath184 .
thus the source has a somewhat directional radiation pattern . ] , but with @xmath185 .
the radiation pattern is increasingly directional and could be produced approximately by a led with an integrated collimating lens . ]
let us consider a simple model for a rotationally symmetric surface - emitting led illustrated in fig .
[ f : ledgeom]a , where the primary light - emitting region is planar ( such as a quantum well ) and buried inside a semiconductor material of refractive index @xmath186 .
each primary source point is assumed to radiate ( independently ) a spherical wave .
if we denote the propagation angle inside the semiconductor material by @xmath187 , the radiant intensity may be expressed as a sum of radially and azimuthally polarized contributions @xmath188 , @xmath9 : @xmath189 where @xmath190 is the complex amplitude of the plane wave in direction @xmath187 .
if the distance between the pn plane and the semiconductor - air interface is large compared to @xmath191 , the local plane wave approximation is valid at the semiconductor - air interface .
thus the output radiant intensity takes the form @xmath192 here @xmath193 , the angles @xmath1 and @xmath187 are related by snell s law @xmath194 , and @xmath195 , where @xmath196 are given by fresnel s equations @xmath197
@xmath198 for radial ( tm ) and azimuthal ( te ) polarizations .
because of the large refractive index @xmath186 of a semiconductor , only a narrow cone of plane waves with incident angles @xmath187 in the range @xmath199 .
we may thus assume that the radial and azimuthal contributions to the radiant intensity of the primary source are equal , i.e. , @xmath200 , and hence we may denote @xmath201 and @xmath202 . then the degree of polarization given by eq
. is @xmath203 and the radiant intensity transforms at the interface according to @xmath204.\ ] ] if we assume that the radiation pattern produced by the primary source is lambertian , with @xmath205 the radial and azimuthal contributions to the radiant intensity of the secondary source become @xmath206 these contributions @xmath207 and @xmath208 are shown by the dotted and dashed lines , respectively , in fig .
[ f : ledgeom]b , where we have taken @xmath209 .
the curves fit well the @xmath0 : we obtain @xmath210 for the azimuthally polarized contribution , @xmath211 for the radially polarized contribution , and @xmath212 for an equally weighted sum of the two contributions .
thus the elementary electric - field modes given by eq .
provide good approximations of the modes of the structure in fig .
[ f : ledgeom]a .
the degree of polarization , also plotted in fig .
[ f : ledgeom]b , increases from a zero on - axis value ( unpolarized radiation in the paraxial domain ) to @xmath213 when @xmath214 , indicating partially polarized radiation in the non - paraxial domain .
is the refractive index of the semiconductor material .
the solid and dashed curves illustrate the azimuthally and radially polarized contributions to the radiant intensity distribution @xmath215 .
( b ) geometrical - optics predictions of the azimuthally ( dashed curve ) and radially ( dotted curve ) polarized contributions to the radiant intensity , their average ( solid curve ) , and the degree of polarization @xmath216 in the far zone ( thick solid curve).,title="fig : " ] is the refractive index of the semiconductor material .
the solid and dashed curves illustrate the azimuthally and radially polarized contributions to the radiant intensity distribution @xmath215 .
( b ) geometrical - optics predictions of the azimuthally ( dashed curve ) and radially ( dotted curve ) polarized contributions to the radiant intensity , their average ( solid curve ) , and the degree of polarization @xmath216 in the far zone ( thick solid curve).,title="fig : " ]
the electromagnetic elementary - mode decomposition presented in this paper should prove useful in in optical system modeling by field tracing methods . to this end , it is necessary to determine elementary field modes and the weight functions of the source .
the example presented above illustrates the possibility of doing this if there is sufficient a priori information about the structure of the source . if , however , the source properties are not known , it is necessary to determine the modal decomposition experimentally . as in the scalar case @xcite
, this can in principle be accomplished by far - field measurements . in general , the polarization basis vectors @xmath217 and the eigenvalues @xmath218 can be determined from the polarization matrix @xmath47 . this matrix can be determined by measuring , e.g. , the angular dependence of the stokes parameters of the field in the far zone .
thus only single - point measurements across the radiation pattern are needed .
determination of the weight functions requires , in general , two - point correlation measurements in the far zone , but this is avoided if the field is known to be quasihomogeneous .
this work was supported by the academy of finland ( 118951 , 129155 , and 209806 ) .
in this appendix we present some details of the derivation of eq .
( [ cylinder ] ) , which is the general representation of the elementary electric - field modes . denoting the circular cylindrical coordinates by @xmath219 in the position - vector space and by @xmath220 in the wave vector space , and
the spherical polar coordinates by @xmath221 in the wave vector space , we have the following relations between the unit vectors of these systems and the cartesian coordinates ( see fig . [ f : notations ] ) : @xmath222 @xmath223 @xmath224 where @xmath225 is the rotation matrix and @xmath134 may stand for @xmath226 , @xmath135 , or @xmath1 . using eq . and recognizing that @xmath227 and @xmath228 , we can write eq . in circular cylindrical coordinates : @xmath229 using eqs .
and one can establish the relation @xmath230 where @xmath231 , and express @xmath97 in the form @xmath232{\rm
e}^{{\rm i}\beta}\left(\boldsymbol{\hat{\rho } } + { \rm i}\boldsymbol{\hat{\psi}}\right)\nonumber\\ & + & \frac{1}{2}\left[\cos\theta f_{j,\theta}(\theta,\psi)+{\rm i } f_{j,\psi}(\theta,\psi)\right]{\rm e}^{-{\rm i}\beta}\left(\boldsymbol{\hat{\rho } } - { \rm i}\boldsymbol{\hat{\psi}}\right)\nonumber\\ & - & \sin\theta f_{j,\theta}(\theta,\psi)\mathbf{\hat{z}}. \label{a6}\end{aligned}\ ] ] recognizing further that @xmath233 and that @xmath234 , we can expand the plane - wave term in eq . in a series form
@xmath235 where @xmath236 denotes the bessel function of the first kind and order @xmath237 .
combining eqs .
and we then have @xmath238 where @xmath239 performing the integration with respect to @xmath135 with the aid of the definition of the azimuthal fourier coefficients in eq . , we obtain @xmath240\nonumber\\ & & \times { \rm e}^{{\rm i}(m+1)\phi } j_m(k\rho\sin\theta)k\sin\theta\nonumber\\ & + & \frac{1}{2}\left(\boldsymbol{\hat{\rho } } - { \rm i}\boldsymbol{\hat{\phi}}\right)\left[\cos\theta f_{j,\theta , m-1}(\theta)+{\rm i}f_{j,\psi , m-1}(\theta)\right]\nonumber\\ & & \times { \rm e}^{{\rm i}(m-1)\phi}j_m(k\rho\sin\theta)k\sin\theta\nonumber\\ & - & \mathbf{\hat{z}}\,{\rm e}^{{\rm i}m\phi } j_m(k\rho\sin\theta)k\sin^2\theta.\end{aligned}\ ] ] collecting terms in the summation of eq . by replacements @xmath241 and employing the bessel - function identities @xcite @xmath242 and @xmath243 we finally arrive at eq . .
let us first recall the series - representation of bessel functions @xcite : @xmath244 on the other hand , we have the relation @xcite @xmath245 where @xmath246 , @xmath247 , @xmath248 is the beta function , and @xmath249 is the gamma function @xcite . combining eqs . and , and interchanging the order of summation and integration yields \10 .
l. mandel and e. wolf , _ optical coherence and quantum optics _ ( cambridge university press , cambridge , uk , 1995 ) .
e. wolf , `` new theory of partial coherence in the space
frequency domain .
part i : spectra and cross spectra of steady - state sources , '' j. opt .
* 72 , * 343351 ( 1982 ) .
f. gori , m. santarsiero , r. simon , g. piquero , r. borghi , and g. guattari , `` coherent - mode decomposition of partially coherent , partially polarized sources , '' j. opt .
am . a * 20 , * 7884 ( 2003 ) .
j. tervo , t. setl , and a. t. friberg , `` theory of partially coherent electromagnetic fields in the space frequency domain , '' j. opt .
. am . a * 21 , * 22052215 ( 2004 ) .
f. gori , `` collett - wolf sources and multimode lasers , '' opt .
* 34 , * 301305 ( 1980 ) .
f. gori , `` mode propagation of the light field generated by collett - wolf schell - model sources , '' opt .
* 46 , * 149154 ( 1983 ) .
a. starikov and e. wolf , `` coherent - mode representation of gaussian schell - model sources and of their radiation fields , '' j. opt .
am . a * 72 , * 923928 ( 1982 ) .
a. starikov , `` effective number of degrees of freedom of partially coherent sources , '' j. opt .
* 73 , * 15381544 ( 1983 ) .
j. huttunen , a. t. friberg , and j. turunen , `` diffraction of partially coherent electromagnetic fields by microstructured media , '' phys
e * 52 , * 30813092 ( 1995 ) .
p. vahimaa and j. turunen , `` bragg diffraction of spatially partially coherent fields , '' j. opt .
am . a * 14 , * 5459 ( 1997 ) .
f. gori and c. palma , `` partially coherent sources which give rise to highly directional laser beams , '' opt
* 27 , * 185188 ( 1978 ) .
f. gori , `` directionality and partial coherence , '' opt .
acta * 27 , * 10251034 ( 1980 ) .
p. vahimaa and j. turunen , `` finite - elementary - source model for partially coherent radiation , '' opt .
express * 14 , * 13761381 ( 2006 ) . f. gori and m. santarsiero , `` devising genuine correlation functions , '' opt .
* 32 , * 531 - 3533 ( 2007 ) .
j. turunen and p. vahimaa , `` independent - elementary - field model for three - dimensional spatially partially coherent sources , '' opt .
express * 16 , * 64336442 ( 2008 ) .
m. a. alonso and e. wolf , `` the cross - spectral density matrix of a planar , electromagnetic stochastic source as a correlation matrix , '' opt .
* 281 , * 23932396 ( 2008 ) .
j. tervo and j. turunen , `` angular spectrum representation of partially coherent electromagnetic fields , '' opt . commun .
* 209 , * 716 ( 2002 ) .
t. setl , j. tervo , and a. t. friberg , `` complete electromagnetic coherence in space frequency domain , '' opt . lett . * 29 , * 328330 ( 2004 )
. j. tervo , t. setl , and a. t. friberg , `` degree of coherence for electromagnetic fields , '' opt .
express * 11 , * 11371143 ( 2003 ) .
g. b. arfken and h. j. weber , _ mathematical methods for physicists _
( academic press , san diego , ca , 2001 ) . i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series , and products _ , a. jeffrey , d. zwillinger , eds . , seventh edition .
( academic press , orlando , 2007 ) . w. n. bailey , _ generalized hypergeometric series _ , vol .
32 of cambridge tracts in mathematics and mathematical physics , g. h. hardy and e. cunningham , eds .
( cambridge university press , cambridge , uk , 1935 ) . | a representation of partially spatially coherent and partially polarized stationary electromagnetic fields is given in terms of mutually uncorrelated , transversely shifted , fully coherent and polarized elementary electric - field modes .
this representation allows one to propagate non - paraxial partially coherent vector fields using techniques for spatially fully coherent fields , which are numerically far more efficient than methods for propagating correlation functions .
a procedure is given to determine the elementary modes from the radiant intensity and the far - zone polarization properties of the entire field .
the method is applied to quasihomogeneous fields with rotationally symmetric @xmath0 radiant intensity distributions ( @xmath1 being the diffraction angle with respect to the optical axis and @xmath2 an integer ) .
this is an adequate model for fields emitted by , e.g. , many light - emitting diodes . |
the configuration - mediated directionality of non - covalent bonds between proteins explains their propensity to self - assemble into fibrils and filaments @xcite .
protein filaments are ubiquitous in biology , forming inside the cells or in the extra - cellular matrix individually , in bundles , or in randomly crosslinked networks .
they facilitate the propulsion in bacteria , they control the mechanical strength in cytoskeleton and the bending stiffness in axons , they allow positional control of organelles and provide the transport routes all around the cell @xcite . in a different situation ,
the self - assembly of proteins into amyloid fibrils impairs physiological activity and is the root cause of a number of organic dysfunctions @xcite . in yet another context
, filaments are artificially or spontaneously assembled to achieve a specific function in the material , such as directed conductivity , plasmonic resonances , or just the mechanical strength in a fiber composite , with important technological applications @xcite .
finally , a conceptually related issue emerges in the denaturation of dna @xcite , for which the available theoretical framework @xcite can not provide predictions about the topology of the disassembly process .
the typical size of all these aggregates , and its time - evolution , are a non - trivial function of the rate at which bonds along the filament spontaneously dissociate due to the thermal motion of the assembled molecules .
the dissociation rate and the distribution of fragments are important parameters which enter the master kinetic equation description of self - assembling filament size and populations .
a filament growth can be summarized by the reversible reaction : @xmath0 , where the monomer subunit @xmath1 is added to an existing filament of @xmath2-units long . for the forward reaction
, it is commonly accepted that association proceeds by the addition of a single subunit as opposed to the joining of larger segments because of the greater abundance of monomers with respect to active fragments .
in contrast , despite the importance of thermal breakup in many fields of colloid science and technology @xcite , its basic understanding is far from satisfactory .
several studies aimed to explain thermally - activated filament breakup in physical terms , came to the conclusion that fibrils of any respective size can aggregate , while the filament breakup can occur with equal probability anywhere along its length .
in particular , lee @xcite has demonstrated that the thermal breakup occurs randomly along the chain , leading to daughter fragments of any size . in yet another classical model based on equilibrium detailed - balance between the various aggregation and breakup events , by hill @xcite , the highest breakup probability is for two fragments of equal size , i.e. the breakup rate is maximum in the middle .
theoretical models in the past have focused on the simplified case of chains of harmonically bonded particles ( subunits ) , so that the binding force is linear in the inter - protein displacement @xcite . in this approximation
the normal modes of vibration of the chain are de - coupled , which makes the problem amenable to simpler analysis . even in this case
, previous theoretical models reached contradictory conclusions , with either flat breakup distribution or a pronounced maximum in the middle .
however , the physical bonds linking protein filament subunits ( such as hydrogen bonds and hydrophobic attraction ) are strongly anharmonic
. then the problem becomes one of coupled nonlinear oscillators as in the famous fermi - pasta - ulam problem @xcite , for which the typical vibration modes are no longer delocalized periodic waves but solitons @xcite .
this is also consistent with the finding @xcite that in a strained lennard - jones chain , the strain is not uniformly distributed , but localized around the bond which is going to break first .
the standard tools of chemical dynamics and stochastic rate theory @xcite , all based on the harmonic approximation and on normal modes , are therefore inapplicable @xcite . here we develop a systematic microscopic understanding of this process based on brownian dynamics simulation and theoretical arguments , focusing on the nonequilibrium breakup phenomena .
hence we study the intrinsic breakup rates independent of any recombination phenomena which may occur at later stages leading eventually to an equilibrium size .
first of all , we discover that the topology of filament breakup critically depends on the bending stiffness of the chain .
secondly , a clear connection is found between the anharmonicity of subunit interaction and the fragment distribution resulting from thermal breakup .
the anharmonic lennard - jones or morse - like binding potential in stiff or semiflexible filaments inevitably leads to a very strong preference for the breakup to occur at chain ends , but recover the uniform , flat fragment distribution in the limit of harmonic ( or any other symmetric ) potential .
importantly , it is not the bare anharmonicity which controls this effect , but , more precisely , the _ asymmetry _ of the bonding potential about the minimum ( larger force for bond compression than for extension ) , which is inherent to the most common anharmonic potentials .
as we will show below , it is precisely the asymmetry in the potential which `` breaks the symmetry '' between dissociation rates at the middle of the filament and at the ends .
those rates are equal only for symmetric potentials like harmonic , and they always differ for asymmetric potentials . in contrast , when the intermolecular interaction is purely of the central - force type , i.e. a fully flexible chain with no bending resistance , a bell - like distribution peaked in the middle is obtained in accord with the prediction of the hill model . these findings can be understood with an argument based on counting the degrees of freedom per particle for the different potentials .
these results provide a fundamental link between the features of intermolecular interaction and the filament breakup rate and topology , and can be used in the future to predict , control and manipulate the filament length distribution in a variety of self - assembly processes in biological and nanomaterials applications .
to model a non - covalently bonded filament we use a coarse - grained model of linear chains of brownian particles ( fig.[fig1]a ) bonded by the truncated - shifted lennard - jones ( lj ) potential , @xmath3 - { u_c},~~{\rm { for } } ~r < { r_c}}\\ { 0 , \qquad { \rm { for } } ~r \ge { r_c } } \end{array } } \right.\ ] ] where @xmath4 is the distance between two neighbor proteins @xmath5 and @xmath6 , @xmath7 is the linear size of the monomer unit , and @xmath8 $ ] .
the parameter @xmath9 + 1\}$ ] is set to maintain a constant well depth equal to @xmath10 , independently of @xmath11 .
the lj potential is inherently anharmonic , except in the close proximity of its minimum
. an alternative could be the morse potential , and we have checked that the results do not change qualitatively with its use .
figure [ fig1]b explains what we mean by truncation : the attractive region stretches up to a distance @xmath12 ( indicated by arrows in the plot and measured in terms of lj length scale @xmath7 ) , while the depth of the potential well is kept independently fixed ( measured by @xmath10 , in units of @xmath13 ) .
the shorter the attraction range , the closer is the potential to its harmonic approximation .
for all the data we use @xmath14 , which well approximates the strength of the most common physical interactions such as hydrogen bonds and hydrophobic attraction .
we also include in our analysis the local bending energy , in the form @xmath15 , where @xmath16 is the angle between the directions of bonds from the particle @xmath5 to the preceding ( @xmath17 ) and the subsequent ( @xmath6 ) subunits .
figure [ fig1]d illustrates the way this effect is implemented by imposing pairs of equal and opposite forces on the joining bonds , providing a net torque on the junction .
it is the same algorithm that is used in , e.g. lammps ` angle - harmonic ' system @xcite . the bending modulus @xmath18 , in units of @xmath19 ,
is directly related to the persistence length of the filament via the standard expression @xmath20 . in the chain , for several values of attractive range , measured by @xmath12 and indicated by arrows in the plot .
( c ) the contrast between a combined potential @xmath21 felt by an inner subunit in the filament , bonded on both sides , and the end - subunit bonded by the regular lj potential .
( d ) scheme of the bond - bending force which opposes changes in the angle between two adjacent bonds by applying couples on each adjacent bond . ]
the dynamics of the chain of subunits is governed by the overdamped langevin equation , @xmath22 where @xmath23 is the vector containing the positions of all molecules , @xmath24 is the friction coefficient , the total potential force acting on a given particle , @xmath25 , has contribution from both the lj and the bending couples , and the gaussian stochastic force defined such that @xmath26 and @xmath27 , according to the fluctuation - dissipation theorem . for numerical integration eq .
( [ a1 ] ) is discretised in the form known as the ermak - mccammon equation @xcite : @xmath28 where @xmath29 is randomly extracted from a normal distribution with zero average and unit standard deviation .
the discrete time step is taken as @xmath30 , where the reduced time uint is defined as @xmath31 , and @xmath32 is the diffusion coefficient . for a typical globular protein ( e.g. lysozyme ) , with diameter@xmath33 nm and diffusion coefficient @xmath34 m@xmath35/s @xcite , we obtain @xmath36 @xmath37s .
therefore @xmath38 ps .
each run is initialized with the equilibrium interparticle distance @xmath39 , as a straight chain ( all @xmath40 ) , corresponding to the minimum of all interaction potentials .
a dissociation event is assumed to take place when one of the bonds exceeds the cut - off length ( @xmath12 ) , i.e. @xmath41 , at which point the simulation is terminated and the location of the rupture recorded . the location of the rupture is recorded . to generate the probability distributions plotted in figs .
[ fig3 ] and [ fig4 ] , @xmath42 independent runs are performed and the normalised breakup probability is calculated as @xmath43 where @xmath44 is the total number of recorded breakup events for the bond @xmath45 . for most data
we have reached @xmath46 ; since the runs are independent , the @xmath44 are binomially ( bernoulli ) distributed and the error bars are estimated as @xmath47/n}}$ ] , which always stayed below 10% of the value for @xmath48 .
subunits bonded with @xmath49 and @xmath50 were initialized from a straight conformation and allowed to fluctuate for @xmath51 ts .
the resulting snapshots , for each value of @xmath18 indicated on the image , show the effect of differing persistence length . ]
figure [ fig3 ] shows the main result of our brownian dynamics simulation : on increasing the bending stiffness of the filament , the highly inhomogeneous normalized probability @xmath48 changes from a bell - shaped distribution peaked in the middle ( reminding of the hill model ) , to a completely opposite shape , with a strong preference for single subunits to dissociate from the ends . and lj parameters @xmath14 , @xmath50 . the effect of changing bending stiffness ( increasing persistence length ) is evident : for the chain with essentially no bending penalty ( lowest @xmath18 ) the distribution of fragment sizes is bell - shaped with a maximum in the middle of the middle of the filament .
stiff chains , with a strong bond - bending penalty , instead , feature a nearly homogeneous , flat distribution of fragment sizes with an increasingly large increase of breakup rate at the ends .
there is a broad range of semiflexible filaments that behave in exactly the same way : as `` stiff '' chains . ]
the conclusion arising from this data is clear : there is a broad range of what one could collectively interpret as ` stiff ' filaments , for which the nature of bond - breaking statistics is exactly the same .
these are with the bending stiffness of @xmath52 , and their behaviour does not differ from the last dataset in fig .
[ fig3 ] ( labelled ` stiff ' ) , corresponding to the strictly 1-dimensional filament where only the motions along the chain were permitted . for these stiff or semiflexible filaments
there is a very strong preference to dissociate a single subunit from the chain ends , which does diminish for less symmetric potentials , as demonstrated by fig .
[ fig4 ] below .
however , as the chain becomes increasingly flexible , the ratio of breaking rates at the ends and in the middle gradually reverses , and for a very flexible chain ( @xmath53 in the plot ) the breakup probability resembles the prediction from the hill model .
one can qualitatively understand this effect : for a stiff filament ( as shown in figs .
[ fig1]a and [ figk ] ) , in order to develop a thermal fluctuation large enough to stretch a bond beyond @xmath12 , a whole sequence of bonded particles has to move in a correlated fashion ; this leads to an effectively harmonic potential acting on the middle particles , and diminishes their breaking rate very significantly . on the other hand ,
as the particles in a flexible chain are free to move perpendicular to the bond axis , this coherent motion is not required and the bond breaking statistics is dominated by the single - bond equilibrium .
most protein filaments are quite stiff .
the f - actin has the quoted persistence length @xmath54 m @xcite , and the insulin amyloid filaments : @xmath55 m @xcite .
interestingly , if one measures @xmath56 in the units of constituent protein size ( as the parameter @xmath7 in our case ) , these very different filaments all have @xmath56 between 3000 and 6000 units .
we therefore choose the bending stiffness modulus @xmath57 in all subsequent analysis , which is within the class of ` stiff ' filaments according to the data in fig .
[ fig3 ] . long , bonded by potentials with @xmath58 and @xmath57 .
the harmonic potential with the same depth has a uniform ( homogeneous ) probability of breakup , while at increasing @xmath12 ( and the degree of potential asymmetry about the minimum ) the ends of the chain are increasingly more prone to single - particle depolymerization .
the ratio of the probability of breaking at the end ( dissociation ) to the fragmentation in the middle @xmath59 for @xmath50 . ]
this distribution of breaking points along the chain is equivalent to the distribution of fragment sizes resulting upon breakup .
figure [ fig4 ] shows how this distribution depends on the nature of physical bond between subunits . as we have seen in the illustration , fig .
[ fig1]b , changing the cutoff distance @xmath12 while keeping the depth of the attractive potential well constant ( @xmath10 ) effectively alters the degree of potential asymmetry : the larger the @xmath12 , the more asymmetric the potential is .
we have also independently tested the breaking statistics in an explicitly harmonic potential of the same depth and curvature at the minimum . in the limit of harmonic chain , we recover a completely uniform ( flat ) distribution of fragments , with a very high accuracy .
this is in agreement with the theory of lee @xcite , who assumed harmonic bonds .
on the other hand , fig .
[ fig4 ] clearly demonstrates that , with increasing asymmetry , the breakup probability @xmath48 displays an increasingly strong preference for depolymerization from the ends . for the highly asymmetric ( and also highly anharmonic ) potential with @xmath50 ,
the breakup probability of the outer bonds is over @xmath60 times larger than the one of the innermost bonds . , and
the parameters of bonding potentials @xmath49 , @xmath61 and @xmath57 . the range of enhanced breakup probability @xmath62 at each end remains constant for all @xmath2 . ] another important result is shown in fig.[fig5 ] , where for a given level of lj potential depth and asymmetry , and stiff filament with @xmath57 , as usual , we study the effect of filament length ( the total number of bonded subunits , @xmath2 ) . it is more difficult to normalise the breakup probability @xmath48 this time , because for longer filaments there are more and more ` plateau values ' of the constant ( low ) breakup frequency in the middle , which participate in the original normalisation by the total number of runs , @xmath42 ( effectively uniformly suppressing the values of @xmath48 and thus masking the characteristic ratio @xmath63 ) .
we therefore chose to scale all datasets by their maximum value of @xmath64 , such that the different curves are comparable .
it is clear that with the filament increasing past @xmath65 there is no further change in the characteristic ratio @xmath63 simply the region of ` chain middle ' becomes extended .
perhaps one may regard this as an effective confirmation of the lee model @xcite , since for very long and very stiff filaments a very large middle section has an effectively harmonic bonding , and therefore uniform breakup rate .
it appears , the range of enhanced probability near the ends is relatively constant , @xmath66 .
shorter filaments have the middle region elevated simply because the two end - effects start overlapping . the finding
that , for stiff filaments with asymmetric interaction potentials , the dissociation rate at the end can be substantially larger than the rate of fragmentation in the middle , may be important in the self - assembly kinetics of actin filaments @xcite . there , and in many other cases , the tendency to depolymerize at the end is amplified by the presence of multiple bonds in the interior of the filament , due to the double - stranded helical structure in the case of actin .
in addition , we studied the probability of the first breakup ( irrespective of its position along the filament ) , upon varying @xmath12 and the filament length @xmath2 , for the case of a stiff filament ( limit of large @xmath18 ) which also approximates the case of a 1d aggregate . from the results plotted in log - linear fashion in fig.[fig6 ]
it is clear that the probability for the chain to fracture depends exponentially on time , @xmath67 , with a characteristic breakup time @xmath68 increasing upon increasing the attraction range @xmath12 ( and the asymmetry of bonding potential with that ) .
the average first - breakup time , irrespective of the location on the chain , is defined by @xmath69 , upon normalizing @xmath70 .
this exponential dependence can be understood from the analysis of the many - particle fokker - planck equation , @xmath71 with the smoluchowski operator defined as @xcite @xmath72}\ ] ] acting on the many - particle probability density @xmath73 , where , in supervector notation , @xmath74 is the set of interparticle coordinates .
the probability as a function of time that all bonds remain within the cutoff at a time @xmath75 , that is , the probability that the chain does not break within a time @xmath75 , is given by @xmath76 .
we shall recall that , in supervector notation , the condition @xmath77 means that _ all _ bond vectors ( relative particle coordinates ) in the chain are within an extension smaller than the cutoff @xmath12 .
furthermore , @xmath78 represents the multi - dimensional potential energy landscape given by the superposition of the lennard - jones potentials acting on pairs of molecules .
the first passage / breakup time probability density is defined as the change of @xmath79 between the time @xmath75 and @xmath80 , and is thus given by @xmath81 . combining these equations , with some manipulations ( see e.g. ref .
@xcite ) , it is possible to show that the first - passage time probability density is exactly equal to @xmath82 .
the mean first - breakup time is then defined as the first moment of the first - breakup time probability density , @xmath83dt } $ ] , which is the same quantity as measured from the exponential fits in simulations . the exponential dependence on time can be understood from the analysis of the many - particle fokker - planck equation , eqs .
( [ 2])-([3 ] ) .
its general solution is @xmath84 , where @xmath85 labels the eigenfunctions @xmath86 and eigenvalues @xmath87 of the many - body operator @xmath88 .
, normalized such that it is equal to unity at @xmath89 , is plotted against simulation time measured in timesteps ( ts ) .
different data sets represent the different attraction range @xmath12 , which is our measure of potential asymmetry .
the fitted lines are all simple exponentials , from which we extract the characteristic rate of the first breakup , @xmath90 . ]
the probability to break in the middle as @xmath91 and the one at the end as @xmath92 , with @xmath93 subunits from each end affected , the total rate can be estimated as @xmath94 , which is the solid line in the plot with only a single fitting normalisation factor : @xmath95 ; the deviations at small @xmath2 are clearly due to the overlapping end effects ( see fig.[fig5 ] ) . ] according to the ground - state dominance principle , the time evolution for long filaments ( @xmath96 ) is dominated by the smallest non - zero eigenvalue @xmath97 , such that , recalling the expression for @xmath98 , the time dependence of the first - breakup probability is given by @xmath99 .
hence the breakup probability is indeed exponential in time with a characteristic frequency - scale given by the smallest finite eigenvalue @xmath97 of the many - body operator @xmath88 .
this result explains the exponential dependence on time of the breakup probability observed in the simulations in fig .
[ fig6 ] . also , combining the expressions for @xmath98 and for @xmath68 , it is possible to show that @xmath100 , which confirms that the ground - state of the many - body fokker - planck equation indeed sets the time scale of breakup .
furthermore , the rate @xmath90 grows roughly linearly with the chain length @xmath2 , which is demonstrated in fig .
this particular dependence @xmath101 arises because the number of escape attempts increases with the chain size .
one can show by means of the standard supersymmetric transformation of the fokker - planck equation into the schrdinger equation @xcite , that @xmath102 is analogous to the quantum ground - state energy of an ensemble of @xmath103 bound states , and the ground state energy is extensive ( @xmath104 ) within the quasiparticle approximation @xcite .
we find a useful representation in a map that covers all of the @xmath18-@xmath12 parameter space to study how the location of first - breakup events along the filament changes upon varying both the stiffness @xmath18 and the cutoff or asymmetry @xmath12 .
the results can be represented as a contour plot for the ratio @xmath105 as a function of @xmath18 and @xmath12 .
the contour plot is shown in fig .
the bottom left corner , corresponding to flexible ( low-@xmath18 ) filaments with short - ranged potential close to harmonic ( low-@xmath12 ) , represents conditions where the filament breaks in the middle and the fragment distribution is bell - shaped , in conformity with hill s model predictions . upon increasing both @xmath18 and @xmath12 at the same time ,
breakup in the middle becomes less favourable and the distribution tends to flatten out .
eventually , for very stiff filaments and asymmetric potentials with large @xmath12 the opposite limit of u - shaped fragment distributions with preferential breakup at the filament ends is recovered .
this occurs in the top - left region of the map in fig .
[ fig8 ] . for symmetric binding potentials close to harmonic ( low @xmath12 : along the @xmath18 axis of the contour plot ) , the bell - shaped distribution persists longer upon increasing @xmath18 , eventually transforming into a flat distribution @xmath106 for stiff filaments . on the other side of the map , where @xmath12 is increased for flexible chains
, the bell - shaped distribution persists for flexible chains up to @xmath107 which corresponds to the lj with no cutoff
. in general , the most dramatic change in the breakup location and fragment - distribution shape occurs along the path of steepest ascent , defined as the path parallel to the gradient of the surface . based on our results , the path of steepest ascent and most dramatic evolution in the breakup topology
is approximately identified by the line @xmath108 . as a function of filament stiffness @xmath18 and the asymmetry parameter @xmath12 .
the bottom - left lagoon ( dark ) represents conditions where filaments break in the middle ( bell - shaped distribution , according to hill @xcite ) , while the upper - right ( light ) region of this map represents conditions where filaments break at the ends ( the u - shaped distribution ) and negligibly in the inner locations .
arrows on the top signify that these geodesic lines extrapolate towards @xmath109 .
arrows to the right indicate that there is little further change past @xmath110 .
the dashed geodesic line marks the @xmath106 condition , separating the regions of bell- and u - shape distributions . ] in figs .
[ fig3 ] and [ fig4 ] we have shown that depending on the relative extent of bond - bending and central forces in the intermolecular interaction , the fragment size distribution can change from a u - shaped distribution in the limit of large bond - bending rigidity , to a bell - shaped distribution with opposite curvature in the limit of a purely central - force lennard - jones potential .
intermediate bending stiffness values yield distributions with shape in between the two limiting cases .
it is first important to understand the microscopic origin of this qualitative difference upon varying the bending stiffness in the intermolecular interaction .
since the flexible chain breakup statistics closely resembles the prediction of hill @xcite , we take a similar approach and consider the fragment - size dependence of the breakup rate within a chemical equilibrium assumption and for the special simplifying case of harmonic bonds .
we have checked that with harmonic bonds the same behaviour trend as in fig .
[ fig3 ] is reproduced , with the only difference that the @xmath48 distribution for the stiff filament is flat ( as indeed proven by lee @xcite ) instead of u - shaped in the case of stiff filaments ( as the last curve in fig .
[ fig4 ] shows ) .
that is , the hill - like bell - shaped @xmath48 is the universal result for fully flexible chains .
the equilibrium constant for a dissociation reaction @xmath111 of a filament @xmath2 into two fragments @xmath112 takes the form : @xmath113 , where @xmath114 is the partition function of fragment @xmath115 .
@xmath116 is the dissociation rate , while @xmath117 is the recombination rate of these two fragments .
the latter can be estimated from the diffusion - controlled collision rate of two linear chains , upon accounting for the diffusion coefficient of the two chains ( kirkwood - riseman approximation @xcite ) and for the encounter efficiency of end - to - end collisions of the two chains . in this way
, the size - dependence was found to be @xmath118 @xcite .
the size - dependence of the dissociation rate ( and hence the fragment size - distribution ) can be obtained by replacing this form for the association rate in the expression for the equilibrium constant , and upon evaluating the fragment - size dependence of the partition functions in the numerator of @xmath119 . from classical statistical mechanics , rigid - body translational degrees of freedom of the chain
contribute to the partition function a factor @xmath120 , and rigid - body rotational degrees of freedom contribute an extra factor @xmath121 , since the overall mass of the filament is @xmath104 .
together these two factors give a partition function @xmath122 .
the vibrational contributions of the monomers in the chain factorise in the partition function , as for a chain of harmonic oscillators , resulting in standard factors of the type @xmath123 , where @xmath124 is the einstein frequency . clearly these factors do not contribute to @xmath119 because the corresponding terms in the numerator and denominator cancel .
a full consideration of the normal modes of the linear chain with free ends , beyond the einstein model , leads to an additional nontrivial size - dependence @xmath125 , for vibrations of harmonic spheres in 1d , and to @xmath126 for vibrations in a flexible 3d chain @xcite . in simple terms , upon increasing the chain length ,
more low - energy modes can be accommodated in the spectrum , which causes the partition function to decrease .
the importance of this effect was first recognized by j. frenkel @xcite in the context of nucleation phenomena .
hence with purely central - force interaction in 3d ( flexible chain ) the overall contribution is @xmath127 .
akin to covalent bonds in molecular physics , the bending stiffness introduces additional degrees of freedom for rotations about the bond symmetry axis , which then leads to an overall dependence @xmath128 .
one should note that with spheres and purely central - force bonds there is no such axis of symmetry for the rotations , and the three translational degrees of freedom per particle suffice to describe the vibrational behavior . including all these considerations ,
the dissociation rate will have a dependence on the fragment sizes given by @xmath129 the exponent @xmath130 , which collects all size - dependent contributions of the partition function , is different depending on whether the interaction is purely central - force , or has a bond - bending stiffness .
for central forces , @xmath131 , whereas with semiflexible or stiff chains one has @xmath132 .
the leading contribution is then @xmath133 , with a pronounced bell - shape peaked in the middle for the exclusively central - force flexible chain , and @xmath134 , leading to a much flatter distribution for a chain with bond - bending penalty .
the fact that the slightly u - shaped distribution observed in simulations for stiff filaments is not recovered by this model should be attributed to the various approximations ( kirkwood - riseman for chain diffusion , detailed balance , etc . )
involved in the model , and also to the harmonic approximation of independent linear oscillators underlying the factorization of partition functions .
this argument , however , explains , qualitatively , that a flatter distribution of fragments is to be expected in the presence of bond - bending , due to the additional rotational degrees of freedom about the stiff intermolecular bond symmetry axis , which is absent with purely central - force interactions
. we can briefly comment on the qualitative predictions of this model for the distribution of breakup fragments in realistic amyloid fibrils .
realistic intermolecular forces which bind proteins in amyloid fibrils crucially depend on both electrostatics and temperature .
we shall start considering the role of electrostatics first .
electrostatic repulsion between two bound proteins in a filament is ubiquitous except for solutions at very high ionic strength .
electrostatic repulsion acts to `` lift up '' the bonding minimum , and it may also contribute an additional small energy barrier to the total interaction @xmath135 , with a maximum @xmath136 co - existing or competing with the new lifted attractive minimum .
we denote the new attractive minimum as @xmath137 . due to the fact that the electrostatic energy decreases with @xmath4 , and the maximum is typically at @xmath138 , the lifting up of the bonding minimum by the electrostatic repulsion is not entirely compensated by the energy barrier ( the new maximum in @xmath135 ) .
hence the total energy to be overcome for the particle to escape from the bonding minimum is @xmath139 .
this consideration points towards a role of electrostatics which promotes breakup , or at least , restructuring into a different morphology where the electrostatic energy density is reduced .
this outcome of our analysis is compatible with recent experimental observations where an increased electrostatic repulsions ( e.g. at lower ionic strengths ) is responsible for fission or scission phenomena of larger compact aggregates into smaller and more anisotropic aggregates @xcite .
our simulations show a crossover from a u - shaped fragment distribution into a bell - shaped distribution upon going from high values of bond - bending stiffness @xmath18 to lower values . in our simulations , @xmath18 is fixed and set independently of @xmath140 , the latter being kept constant throughout at varying @xmath18 . in reality , however , @xmath18 and @xmath140 may not be decoupled for a realistic model of amyloid fibrils .
the reason is that the inter - protein bending stiffness @xmath18 originates , microscopically , from the strength of @xmath141-sheets which bind two adjacent proteins in the fibril .
the mechanism is known : due to the planar , sheet - like , nature of two hydrogen - bonded @xmath141-sheets , there is an intrinsic bending resistance against sliding or rolling of the two proteins past each other . the same mechanism provides bending rigidity when two surfaces bonded by many anchored central - force springs are displaced tangentially apart . upon increasing @xmath140 ,
the hydrogen and hydrophobic bonds which keep the two @xmath141-sheets together start to dissociate , leading to lower bending stiffness and lower values of @xmath18 . hence , based on our simulation results , we can predict that the fragment distribution function of realistic amyloid fibrils should evolve from a u - shaped distribution at low temperature @xmath140 , where the @xmath141-sheets of two adjacent proteins are tightly bound , into a bell - shaped distribution at higher @xmath140 where the @xmath141-sheet bonding becomes looser , which makes the bending stiffness @xmath18 decrease .
this prediction seems to be confirmed by preliminary experiments @xcite , and future work using ab - initio simulations should focus on identifying the relationship between @xmath18 and @xmath140 , which controls the evolution of the fragment distribution with @xmath140 . in future research
it will be important to combine all these effects into a general coarse - grained approach along the lines of @xcite , to achieve a bottom - up description of realistic filaments and their size evolution .
when the bending rigidity of the chain is high , the probability of spontaneous bond breaking is flat when the bond potential is harmonic @xcite yet it adopts a very distinct and very strongly biased u - shape when the /asymmetry of the potential increases ( fig.[fig4 ] ) .
how can we quantitatively explain why the asymmetry of interaction potential between any two bonded subunits leads to higher breakup rates at the chain ends , and much smaller breakup rates in the middle ? for a high bending modulus one can treat the bond at the filament end as a classical diatomic molecule , and a subunit in the middle of the chain as the inner particle in a linear triatomic molecule . in the latter case
, the combined potential @xmath142 felt by the particle in the middle is sketched in fig.[fig1]c .
one would be tempted to explain the difference between the higher dissociation rate at the filament end and the lower one in the middle by referring to the overall lower energy ( deeper potential well ) felt by the particle in the middle sitting in the minimum of the combined potential @xmath21 . applying a kramers - type escape - rate argument
would then lead to an arrhenius dependence of the particle on the depth of the energy well and an overall large difference between the two rates .
however , such an approach can not explain the observation that the rate is the same in the middle and at the end for the case of harmonic potential ; in that case the same argument about @xmath142 applies hence one would expect a lower rate in the middle , which is not observed , in agreement with previous calculations @xcite .
what is different in the case of the harmonic potential , is the fact that the asymmetry of the bonding potential is removed for the particle at the end of the chain ( while the subunits in the middle effectively experience the harmonic potential in both cases ) .
it is in fact this asymmetry which facilitates dissociation at the termini of the chain , where less resistance is encountered by the particle escaping outwardly from the bound state . in order to verify that this is indeed the right physics
, we also run a test simulation with a quartic potential @xmath143 , which is anharmonic yet fully symmetric about the minimum , just like the harmonic potential .
also in this case we found a completely flat distribution of fragments , as for the harmonic potential , which supports the proposed claim .
it is therefore the asymmetry , in the case of anharmonic potentials , which plays the major role in facilitating the preferential bond breakup at the chain ends .
the explanation can be found in the different values of the mean thermal fluctuation from the equilibrium position ( energy minimum ) for the particle sitting in the asymmetric lj potential at the chain end , and the particle moving in the more symmetric combined potential @xmath21 in the middle of the chain .
an analysis of the mean thermal fluctuation done long ago by j. frenkel @xcite , shows that the mean thermal fluctuation of the particle feeling the anharmonic / asymmetric potential at the end is typically larger because of the shallower slope of the potential in the outward direction . for the particle in the middle ,
the situation is different because the combined potential @xmath21 does not become shallower as the particle in the middle moves away from one of the two neighbours , due to the presence of the interaction with the other neighbour .
by means of brownian dynamics simulations , we have shown that thermal breakup rates and breakup topology of model protein filaments ( and other linear nanoparticle aggregates ) are strongly affected by the presence of bond - bending stiffness in the interaction between subunits , and by the degree of asymmetry of the anharmonic binding potential . with stiff chains bonded by inter - particle forces with anharmonicity and asymmetry of the potential typical for intermolecular interaction potentials ( van der waals , hydrophobic attraction etc ) ,
we find a strongly preferential breakup at the chain ends , and an overall u - shaped fragment distribution .
in contrast , with purely central - force interactions between subunits , that is , fully flexible chains the fragment size distribution is bell - shaped , with a pronounced peak in the middle ( symmetric breakup ) , and the lowest breakup rate is found at the ends of the chain . while the preferential breakup at the end of stiff chains ( filament depolymerization ) can be explained in terms of the larger thermal fluctuations at the chain - end associated with potential anharmonicity / asymmetry in a perfectly stiff quasi-1d chain model , the dramatic change of breakup topology upon varying the strength of bond - bending interaction
is more subtle .
in this case we found a tentative explanation upon considering the degrees of freedom associated with the vibrational partition function of the fragments .
in general , breakup into two equal fragments is favoured with purely central - force bonds because the product of the partition functions of two fragments is maximised ( which is intuitive if one considers that the classical partition function for rigid body motions increases strongly with the fragment size ) .
the vibrational partition function , instead , decreases with fragment size because more low - energy modes can be accommodated in longer fragments .
this effect becomes stronger in the case of bond - bending , where the total number of vibrational degrees of freedom is larger due to the rotation axis of the stiff bond . as a result of this compensation between the size dependencies of the vibrational and rigid - body partition functions ,
the size - dependence of fragmentation rate with bond - bending is much weaker compared to the central - force case .
hence , we found some general laws which govern the fragmentation behavior of model linear aggregates , as a function of the relative importance of central - force and bond - bending interactions between subunits .
these findings are important towards achieving a bottom - up control over the length and time - evolution of filament populations , both in biological problems ( acting , amyloid fibrils etc . ) and in nanoparticle self - assembly for photonic applications .
we are grateful for many discussions and input of t.p.j .
knowles , t. michaels , c.m .
dobson and a. bausch .
this work has been supported by the ernest oppenheimer fellowship at cambridge ( az , ld ) and by the technische universitt mnchen institute for advanced study , funded by the german excellence initiative and the eu 7th framework programme under grant agreement nr .
291763 ( az ) .
ld also acknowledges the marie curie itn - comploids grant no .
234810 .
f. oosawa , f. s. asakura , _ thermodynamics of the polymerization of protein .
_ ( academic press , 1975 ) .
t. d. pollard , ann . rev
. biochem . * 55 * , 987 - 1035 ( 1986 ) .
f. chiti , c. m. dobson , annu .
biochem . * 75 * , 333 ( 2006 ) .
d. chandler , nature * 437 * , 640 - 647 ( 2005 ) .
a. irbck , s. a. jnsson , n. linnemann , b. linse , s. wallin .
lett . * 110 * , 058101 ( 2013 ) .
p. s. niranjan , p. b. yim , j. g. forbes , s. c. greer , j. dudowicz , k. f. freed , j. f. douglas , j. chem . phys .
* 119 * , 4070 - 4084 ( 2003 ) .
j. adamcik , j .-
jung , j. flakowski , p. de los rios , g. dietler , r. mezzenga , nature nanotech . * 5 * , 423 - 428 ( 2010 ) .
m. tanaka , s. r. collins , b. h. toyama , j. s. weissman , nature * 442 * , 585 ( 2006 ) .
a. zaccone , d. gentili , h. wu , m. morbidelli , j. chem .
132 , 134903 ( 2010 ) .
h. wu , a. tsoutsoura , et al .
langmuir 26 , 2761 ( 2010 ) .
t. p. j. knowles , c. a. waudby , g. l. devlin , s. i. a. cohen , a. aguzzi , m. vendruscolo , e. m. terentjev , m. e. welland , c. m. dobson , science * 326 * , 1533 - 1537 ( 2009 ) .
v. fodera , a. zaccone , m. lattuada , a. m. donald , phys .
lett . * 111 * , 108105 ( 2013 ) .
l. di michele , e. eiser , v. fodera , j. phys .
lett . * 4 * , 3158 ( 2012 ) .
s. odenbach , ed .
_ colloidal magnetic fluids . basics , development and applications of ferrofluids_. ( berlin , springer , 2009 ) .
b. bonn , h. kellay , m. prochnow , k. ben - djiemiaa , j. meunier , science * 280 * , 265 - 267 ( 1998 ) .
l. dimichele , et al .
136 , 6538 ( 2014 ) .
m. peyrard , a. r. bishop , phys .
. lett . * 62 * , 2755 - 2758 ( 1989 ) . c. b. mast , s. schink , u. gerland , d. braun , proc .
usa * 110 * , 8030 - 8035 ( 2013 ) . c. f. lee , phys .
e * 80 * , 031134 ( 2009 ) . t. l. hill ,
j. 44 , 285 ( 1983 ) .
e. fermi , j. r. pasta , s. ulam , los alamos scientific laboratory report no .
la-1940 , may 1955 .
n. j. zabusky , m. d. kruskal , phys .
lett . * 15 * , 240 - 243 ( 1965 ) .
f. a. oliveira , p. l. taylor , j. chem.phys .
* 101 * , 10118 ( 1994 ) .
a. ghosh , d. i. dimitrov , v. g. rostiashvili , a. milchev , t. a. vilgis , j. chem .
* 132 * , 204902 ( 2010 ) .
a. zaccone , e. m. terentjev , phys .
rev . lett .
* 108 * , 038302 ( 2012 ) .
p. haenggi , p. talkner , m. borkovec , rev .
* 62 * , 251 - 341 ( 1990 ) .
f. a. l. mauguiere , p. collins , g. s. ezra , s. wiggins , j. chem . phys . * 138 * , 134118 ( 2013 ) .
j. paturej , a. milchev , v. g. rostiashvili , t. a. vilgis , j. chem . phys . * 134 * , 224901 ( 2011 ) .
s. j. plimpton , j. comput . phys .
* 117 * , 1 ( 1995 ) .
d. l. ermak , j. chem . phys . * 62 * , 4189 ( 1975 ) .
d. l. ermak , j. a. mccammon , j. chem . phys . * 69 * , 1352 ( 1978 ) .
d. burne , s. kim , proc .
natl . acad .
90 * , 3835 ( 1993 ) .
t. yanagida , m. nakase , k. nishiyama , f. oosawa , nature , * 307 * , 58 - 60 ( 1984 ) .
f. gittes , b. mickey , j. nettleton , j. howard , j. cell biol . *
120 * , 923 - 934 ( 1993 ) .
t. p. j. knowles , j. f. smith , a. craig , c. m. dobson , m. e. welland , phys .
lett . * 96 * , 238301 ( 2006 ) .
j. riseman , j. g. kirkwood , j. chem . phys . * 18 * , 512 ( 1950 ) .
f. f. abraham and j. canosa , j. chem .
50 * , 1303 ( 1969 ) .
j. lothe and g. m. pound , phys . rev .
* 182 * , 339 ( 1969 ) .
j. frenkel , _ kinetic theory of liquids _
( dover , new york , 1946 ) .
m. doi , s. f. edwards .
_ the theory of polymer dynamics _ ( oxford university press , 1986 )
. n. g. van kampen .
_ stochastic processes in physics and chemistry _
( elsevier , amsterdam , 1997 ) .
w. ebeling and i.m .
sokolov , _ statistical thermodynamics and stochastic theory of nonequilibrium systems _ ( world scientific , singapore , 2005 ) .
d. pines and p. nozieres , _ the theory of quantum liquids _ , vol . 1 ( w.a .
benjamin , reading massachusetts , 1966 ) .
a. dehsorkhi , v. castelletto , i. w. hamley , j. adamcik , r. mezzenga , soft matter 9 , 6033 - 6036 ( 2013 ) .
l. nicoud , s. lazzari , d. balderas barragan , and m. morbidelli , preprint ( 2015 ) .
knowles , et al .
109 , 158101 ( 2012 ) .
s. assenza , j. adamcik , r. mezzenga , p. de los rios , phys .
113 , 268103 ( 2014 ) . | protein molecules often self - assemble by means of non - covalent physical bonds to form extended filaments , such as amyloids , f - actin , intermediate filaments and many others .
the kinetics of filament growth is limited by the disassembly rate , at which inter - protein bonds break due to the thermal motion .
existing models often assume that the thermal dissociation of subunits occurs uniformly along the filament , or even preferentially in the middle , while the well - known propensity of f - actin to depolymerize from one end is mediated by chemical factors ( adp complexation ) . here
we show for a very general ( and generic ) model , using brownian dynamics simulations and theory , that the breakup location along the filament is strongly controlled by the asymmetry of the binding force about the minimum , as well as by the bending stiffness of the filament .
we provide the basic connection between the features of the interaction potential between subunits and the breakup topology . with central - force ( that is , fully flexible ) bonds
the breakup rate is always maximum in the middle of the chain , whereas for semiflexible or stiff filaments this rate is either a minimum in the middle or flat .
the emerging framework provides a unifying understanding of biopolymer fragmentation and depolymerization , and recovers earlier results in its different limits . |
spin - exchange interactions are of fundamental importance for the magnetic properties of strongly correlated media @xcite .
of particular interest are superexchange interactions that are induced by higher - order tunneling processes via virtually occupied intermediate states @xcite . in solid state systems such as cuo and mno ,
theses processes are of long - ranged nature , as they do not depend on a direct wave - function overlap of the electrons in contrast to direct exchange interactions . in the context of high - temperature superconductivity they are in the focus of current investigations @xcite .
experiments with ultracold atoms in optical lattices present a promising tool for the study of fundamental properties of higher - order tunneling processes @xcite . here , typically second - order superexchange tunneling is considered .
however , also models with third - order tunneling processes as the leading order are discussed in the literature , e.g. , strongly correlated fermions in frustrated lattices @xcite .
recently , the direct time - resolved observation of superexchange processes has been demonstrated in experiments with ultracold atoms in deep double - well potentials @xcite .
these experiments exploit the interaction blockade inhibiting first - order tunneling processes . however , for the interaction blockade the spin fluctuations arising from superexchange are diminished due to the small amplitude .
there is great current interest in accomplishing long - ranged anti - ferromagnetic ordering with ultracold fermions in optical lattices .
however , the small energy associated with superexchange can not compete with thermal fluctuations in present experiments with isotropic lattices @xcite .
the realization of fermionic high - spin systems with ultracold atoms promises a deeper understanding of fundamental spin - spin interactions @xcite .
these systems are of broad interest , since they offer the possibility to realize , e.g. , a multitude of new quantum phases @xcite , the simulation of su(n ) magnetism @xcite as well as spin changing collisions @xcite .
recent experimental advances demonstrate long - lived coherent spin dynamics of fermionic @xmath1 atoms in optical lattices with tunable effective spin and reliable initial - state preparation @xcite .
the interplay of spin - changing collisions and tunneling processes leads to new phenomena such as the instability of an initially band insulating state @xcite .
this allows for the detailed study of many - body non - equilibrium dynamics in a highly controllable environment . here
, we investigate the non - equilibrium superexchange dynamics of fermionic high - spin systems with spin - changing collisions in optical double - well and plaquette lattices .
we find a regime dominated by superexchange with large amplitude and strong spin fluctuations in _ shallow _ potentials .
usually , the observation of superexchange with ultracold quantum gases requires deep lattices with a large interaction gap @xmath2 that suppresses the first - order tunneling @xmath3 . in the latter regime ,
the superexchange has small amplitudes @xmath4 and is therefore extremely sensitive to thermal fluctuations . in a high - spin system ,
spin - changing collisions lift the pauli blocking of an initially prepared band - insulator state with @xmath5 .
this causes the initial state to be unstable which in general allows strong particle - number fluctuations to arise @xcite .
surprisingly , we find that the first - order tunneling is strongly suppressed in shallow few - well potentials . in this regime , an emerging tunneling- energy gap leads to an interaction blockade at large @xmath6 .
we show how this phenomenon can be used to investigate superexchange processes with large amplitude corresponding to short time scales .
this opens new possibilities for the dynamical study of superexchange in a non - equilibrium system , previously restricted to deep optical lattices with small superexchange amplitude @xmath4 .
we present a comprehensive study of the exact time evolution of a band insulator in double - well potentials and four - well plaquettes . a possible experimental realization with @xmath1 and a measurement scheme for particle - number fluctuations
is discussed .
the required techniques , i.e. , the preparation of optical few - well potentials @xcite and the reliable initial state preparation @xcite have already been successfully demonstrated .
system , a spin - changing collision transfers particles on a lattice site from @xmath7 to @xmath8 and vice versa .
( b ) a superexchange process exchanges two particles on neighboring lattice sites . here
the orange @xmath9 particle on the left site is exchanged with the green @xmath10 particle on the right site .
( c ) a first - order tunneling processes inducing particle - number fluctuations .
[ fig_schemes],scaledwidth=50.0% ] and the blue line the spin fluctuations @xmath11 averaged over a long period of time ( @xmath12 ) as a function of @xmath6 .
the dashed lines shows the fluctuations @xmath13 of the respective ground - state of the system for comparison .
the shading illustrates the crossover between the two regimes @xmath14 and @xmath15 .
( b ) the same data evaluated for the time evolution in a four - well plaquette .
additional to the two limiting regimes , sharp resonances with high fluctuations occur for specific values of @xmath6 .
[ fig_fluctuations],scaledwidth=90.0% ] in this paper we discuss the realization of a pure superexchange model with high - spin fermions .
first , we introduce the corresponding superexchange as well as the full microscopic hamiltonian . in the subsequent sections ,
we demonstrate that the superexchange hamiltonian @xmath16 can be realized with particles of spin @xmath17 in an optical lattice in a surprisingly wide range of parameters , including shallow lattices with large amplitudes @xmath4 . here , @xmath18 ( @xmath19 ) creates ( annihilates ) a particle on lattice site @xmath20 in the spin - state @xmath21 and obeys the fermionic commutation relations .
the first part of this hamiltonian describes spin - changing collisions with an amplitude @xmath22 as depicted in figure [ fig_schemes](a ) .
two particles on a lattice site exchange their spins while preserving the total magnetization , i.e. , @xmath23 . the second part of the hamiltonian represents superexchange processes ( figure [ fig_schemes](b ) ) .
these second - order tunneling processes exchange particles on neighboring lattices sites , which preserves the density distribution but can lead to strong spin - fluctuations already in a double - well system .
the corresponding amplitude @xmath4 accounts for the increased interaction energy @xmath2 in the virtually occupied intermediate state .
the interaction energy itself is omitted in as a constant offset , since neither spin - changing collisions nor the superexchange affects the particle - number distribution .
the first - order tunneling ( figure [ fig_schemes](c ) ) is not present in the hamiltonian and any intersite dynamics is mediated by superexchange processes .
a system of fermionic particles with spin @xmath17 in an optical lattice can be used to realize the hamiltonian .
the full microscopic description is given by the hamiltonian @xmath24 the number operator @xmath25 counts particles on a lattice site @xmath20 .
in contrast to the superexchange hamiltonian , the full model allows for first - order tunneling @xmath26 , e.g. , @xmath27 ( figure [ fig_schemes](c ) ) and a corresponding change in onsite - interaction energy @xmath2 .
the spin - changing collisions represent a small perturbation of the standard hubbard hamiltonian @xmath28 .
we assume an amplitude of @xmath29 , which is a realistic value for @xmath1 ( see section [ sec_spin - dependent ] ) . in this work , we investigate the dynamics of the band insulating ground state @xmath30 of the unperturbed hamiltonian @xmath31 .
the robust experimental preparation of this two - component band insulator as well as the control over the spin - changing collisions @xmath32 via a magnetic field has been demonstrated in reference @xcite . in the following
we discuss the dynamics of in a double - well and a four - well plaquette potential .
both setups can be realized experimentally @xcite and can be solved numerically exact .
we show that the time evolution is governed by the superexchange model even in the absence of a direct interaction blockade .
this effect depends on the existence of the fluctuationless eigenstate of the unperturbed hamiltonian in shallow lattices .
note that there is no such eigenstate in bosonic systems where the realization of a pure superexchange model is only possible in deep lattices .
by superimposing two optical lattices of a short wavelength @xmath33 and a longer wavelength @xmath34 an array of double - well potentials can be realized @xcite .
the preparation of the initial state can be achieved by loading atoms with @xmath5 into the optical lattice of the short wavelength before ramping up the second lattice to separate the individual double - wells .
the time evolution can be determined by means of exact diagonalization of the microscopic hamiltonian . for a wide range of parameters @xmath6
we investigate the dynamics and evaluate the time - dependent particle - number fluctuations @xmath35 . in figure [ fig_fluctuations](a ) the time - averaged value of @xmath36 is shown as a function of @xmath6 ( solid black line ) and compared with the corresponding value of the ground state ( dashed line ) . for deep lattices , fluctuations are exponentially suppressed for decreasing @xmath6 due to an interaction blockade @xmath2 .
surprisingly and in contrast to the ground state , we also find an exponential suppression of @xmath36 in shallow lattices for increasing @xmath6 .
therefore , the time evolution of the band insulator is governed by the superexchange hamiltonian for almost all lattice depths .
this phenomenon allows the simulation of the superexchange hamiltonian with ultracold atoms without an interaction blockade , i.e. , with large amplitudes @xmath4 . in the case of vanishing tunneling @xmath37 , i.e. , in a very deep lattice , both sites undergo in - phase rabi oscillations induced by the spin - changing collisions ( figure [ fig_double - well](a ) ) .
the state remains fluctuationless due to a large interaction blockade @xmath2 . after an evolution time of @xmath38
, the state will eventually be the band - insulator @xmath39 . in both states ,
@xmath40 and @xmath41 , no tunneling is possible due to the pauli principle . for all other times , however , the state is in a superposition of @xmath42 and @xmath43 on both wells . for a finite amplitude @xmath44
first - order tunneling is in principle possible ( figure [ fig_schemes](c ) ) .
surprisingly , the numerical results show very low occurrence of particle - number fluctuations @xmath36 even and especially for large values of @xmath6 .
thus , the dynamics of is described by the superexchange operator .
fig_double - well ] shows the time evolution for various parameters @xmath6 calculated with the full hamiltonian ( solid lines ) and with the superexchange hamiltonian ( dotted lines ) . only at values of @xmath45 , where @xmath36 has a maximum , finite deviations between the two models are present ( figure [ fig_double - well](b ) ) . in shallow lattices ,
periodic oscillations with full amplitude are reestablished ( figure [ fig_double - well](c ) ) .
superexchange processes take place on a short time - scale , leading to the population of onsite states with finite magnetization @xmath46 such as @xmath47 ( black lines in figure [ fig_double - well](b , c ) ) .
this directly corresponds to a non - zero value of spin fluctuations @xmath48 , whereas particle - number fluctuations @xmath36 are negligible .
in contrast , in deep lattices spin fluctuations remain small . here , @xmath11 is on the order of @xmath49 and thus the black line in figure [ fig_double - well](a ) is not visible .
the time - averaged value of the spin fluctuations in the evolution of is shown as a blue line in figure [ fig_fluctuations](a ) .
we can identify two distinct regimes : one with almost no spin fluctuations @xmath50 for small @xmath6 and one with large spin fluctuations @xmath51 for large @xmath6 .
the crossover between the two distinct regimes discussed above is not a unique feature of the double - well but occurs also in other few - well systems . in the following
we identify the underlying mechanism of the fluctuationless time - evolution in shallow potentials for the particular case of a four - well plaquette .
this mechanism is , however , general for all few - well systems that show the discussed behavior .
a four - well plaquette is the natural extension of a double - well potential in two dimensions .
it is equivalent to a one - dimensional four - well lattice with periodic boundary conditions and has been recently realized experimentally @xcite . despite the increased complexity compared with the double - well , the hamiltonian can still be diagonalized for a four - well lattice allowing to investigate the time - evolution of the initially prepared band - insulator exactly .
the hamiltonian has a number of symmetries that can be exploited to reduce the numerical effort .
these symmetries are the spatial translation , spin - flip , quadrupole and particle - hole symmetry .
furthermore , the initial state also obeys several symmetries and we can restrict the basis to the corresponding subspace .
we investigate a wide range of parameters and find the two regimes with @xmath52 , known from the double - well .
the spin fluctuations @xmath11 that arise from the initial state and the ground state properties of the system ( figure [ fig_fluctuations](b ) ) are very similar to the double - well system .
in contrast to the results of the double - well , however , strong fluctuations appear at certain values of @xmath6 .
the eigenspectra allow for a qualitative understanding of the two limits @xmath37 and @xmath53 and also explain the emerging resonances with high fluctuations @xmath36 . in deep potentials
@xmath37 , the onsite interaction @xmath2 is the dominating energy scale of the hamiltonian and the eigenstates separate in groups of fock states with integer @xmath2-excitations that are perturbed by @xmath26 and @xmath32 .
the spectrum is shown in figure [ fig_plaquette](b ) , where each marker corresponds to an eigenstate @xmath54 with an energy @xmath55 and particle - number fluctuations @xmath36 .
the latter directly correspond to the interaction energy @xmath56 @xmath57 red markers have finite overlap with the initial state , which is clearly confined to the group of eigenstates without fluctuations . here
, the interaction blockade @xmath58 causes the dynamics to be determined by the superexchange hamiltonian , which is used in conventional realizations with ultracold atoms @xcite .
the time - evolution of the initial state is shown in figure [ fig_plaquette](a ) , where the solid lines correspond to the full model and the dotted lines to the superexchange model .
note that due to the very small amplitude of the superexchange @xmath4 spin fluctuations @xmath11 are suppressed and merely cause a slow dephasing of the onsite rabi oscillations . in very shallow potentials ,
the tunneling operator @xmath26 is the dominating energy scale of the hamiltonian .
the spectrum for @xmath59 is shown in figure [ fig_plaquette](f ) . in a finite lattice
, the spectrum is gaped due to the discrete energies of the bloch functions . in the case of four lattice sites and taking into account the conservation of momentum , the gap is @xmath60 .
when this is large compared with @xmath2 and @xmath22 , the spectrum separates in groups with @xmath61 where @xmath62 is an integer number .
the initial state obeys @xmath63 and @xmath64 and consequently lies in the group with @xmath65 . for @xmath66 , in this group the eigenenergies are given by the interaction energy @xmath67 , similar to the regime of the interaction blockade , as indicated by the dashed lines in the spectra .
since the spin - changing collisions present only a small perturbation @xmath68 , the band isolator state is only perturbed by states with @xmath69 . for these states ,
the particle - number fluctuations are small and the band isolator state can only develop small particle - number fluctuations .
interestingly , a subordinate energy gap @xmath70 arises between fluctuationless states with @xmath69 ( including @xmath40 ) and states with particle - number fluctuations and an energy @xmath71 . due to the tunneling - energy gap
, the interaction blockade is reestablished in sufficiently shallow potentials , where now @xmath22 is the competing energy scale instead of @xmath3 .
the gap @xmath58 amplifies the suppression of particle - number fluctuations for the initial state , but is not a necessary requirement for @xmath72 .
the surprisingly low fluctuations in very shallow potentials thus arise from a tunneling - energy gap @xmath73 and a subordinate interaction blockade . in figure [ fig_plaquette](e ) the time evolution for @xmath59
is shown , where no deviations between the full model and the superexchange model can be seen .
the time evolution of figure [ fig_plaquette](e ) in units of @xmath74 is identical for all lattice depths with @xmath75 and sufficiently small fluctuations @xmath36 ,
i.e. , away from resonances in figure [ fig_fluctuations](b ) . in the parameter regime @xmath76 ,
neither an interaction blockade @xmath58 nor a tunneling - energy blockade @xmath73 inhibits particle - number fluctuations ( see figure [ fig_plaquette](d ) ) .
the eigenenergies are not determined by the interaction energy only and thus do not directly correspond to the particle - number fluctuations .
however , few eigenstates have finite overlap with the initial state and thus , only at specific parameters @xmath6 fluctuations arise . at these resonances ,
the time evolution of the full model and the superexchange model differ significantly , as can be seen in figure [ fig_plaquette](c ) .
the superexchange model ( dotted line ) predicts the same behavior as for very shallow potentials , whereas the correct calculations are strongly affected by the occurrence of triply occupied sites , which is plotted as a dashed blue line .
, where the interaction - energy shifts have to be taken into account .
( b ) the bias can be tuned to fulfill the resonance condition for certain particle - number distributions @xmath77 , allowing for the measurement of triply occupied wells ( blue line ) .
[ fig_measure_triplons],scaledwidth=65.0% ] an experimental proof of superexchange processes and the absence of first - order tunneling is the demonstration of strong spin fluctuations and vanishing particle - number fluctuations .
the former can be accessed in an experiment with @xmath1 via the population of new spin components , as discussed in section [ sec_spin - dependent ] .
the latter can be measured in an optical double - well setup with the technique illustrated in figure [ fig_measure_triplons](a ) .
first , the lattice with the short wavelength @xmath33 is ramped up to freeze the populations . afterwards
, the relative phase between the two lattice beams is tuned to introduce a bias @xmath78 between the wells . that way the third vibrational mode of the deeper well can be brought into resonance with the first one of the shallower well .
the required bias @xmath79 between the left ( l ) and the right ( r ) well depends on the single - particle energies @xmath80 of the vibrational modes @xmath20 and the particle - number distribution @xmath77 that introduces interaction energy shifts .
when a tunneling process occurs , the interaction energy on the left well is reduced by @xmath81 , while it is increased on the right well by @xmath82 , where @xmath83 is given by the wave - function overlap of the lowest mode with the third one .
this technique has already been successfully demonstrated in reference @xcite .
for every given particle - number distribution we schematically plot the resonance of a transfer to the third band as a function of the applied bias @xmath78 in figure [ fig_measure_triplons](b ) .
the blue line corresponds to the case of three particles in the left and one in the right well .
we assume a width of @xmath84 for the resonances , which is a pessimistic assumption compared with the measured width in the experiment of reference @xcite .
figure [ fig_measure_triplons](b ) shows that it is clearly possible to tune the bias to a value such that particles will be transferred to the third band only if there are exactly three particles on the shallower site .
the population of the third mode can be measured by means of the band - mapping technique afterwards .
as a concrete example , we now consider @xmath1 prepared in the @xmath85 manifold as a possible species for an experimental realization of high - amplitude superexchange . as the initial state
we suggest a mixture of @xmath86 and @xmath87 , which allows for rich spin - dynamics and is the most similar to a band insulator in a @xmath88 system . without tunneling , i.e. , in very deep lattices
, rabi - oscillations between @xmath89 and @xmath90 occur in phase on every lattice site in analogy to the @xmath88 system .
when tunneling is allowed , however , more complicated dynamics arise .
consecutive to tunneling processes , new spin - components can be populated by other spin - changing collisions , e.g. @xmath91 .
states with @xmath92 are only occupied by third - order processes and we neglect them in our simulations as their influence on the qualitative behavior is assumed to be small .
this reduces the system effectively to 6 spin components .
the spin fluctuations @xmath11 are no directly measurable quantity , however the population of the new spin components is proof for the occurrence of tunneling or superexchange processes .
another difference compared with the spin-3/2 system is that the interaction energies @xmath2 and @xmath22 are no longer spin - independent but slightly differ from each other @xcite .
the generalization of the microscopic hamiltonian to 6 spin components is straight forward , when the spin - dependent interaction energies are known . for our simulations we use parameters from @xcite . as a characteristic quantity for the potential depth
we use the ratio @xmath93 , where the interaction energy @xmath94 corresponds to a spin - preserving collision in the initial state .
the energy @xmath95 of the spin - changing collision @xmath96 is significantly smaller with @xmath97 .
the quadratic zeeman - energy must be used to tune the spin - changing collisions into resonance , because the initial and final state of a spin - changing collision have different interaction energies . of course
, this can not be done for all spin states at the same time and the exact time - behavior depends on the applied magnetic field .
our simulations are performed with @xmath98 , but any other value near the resonance is suitable . in figure [ fig_6_components](a , b )
the time - behavior of the @xmath99-components is plotted for the double - well and the plaquette .
they differ from the behavior of a @xmath88 system with spin - independent interaction energies in several ways . in the case of spin - dependent interactions
the superexchange operator @xmath100 can be approximated by the matrix elements @xmath101 between an initial state @xmath102 and a final state @xmath103 with exactly two particles per lattice site , respectively .
the sum is over all possible intermediate states @xmath104 and @xmath105 . for flat lattices ,
the time - evolution can be described by this operator as can be seen by comparing the dotted lines with the solid lines in figure [ fig_6_components](a , b ) .
the new spin - components @xmath106 and @xmath107 occur after a few milliseconds , which corresponds to an increase in spin fluctuations @xmath11 . in the plaquette
the coherent rabi - oscillations are damped out rapidly in strong contrast to the spin-3/2 system , where they last for arbitrarily long times .
the reason for this is the larger number of different frequencies due to the spin - dependency of the interaction strengths . concerning the appearance of particle - number fluctuations @xmath36 , however , the qualitative behavior is identical to the spin-3/2 system , as shown in figure [ fig_6_components](c ) .
the population of the components @xmath10 and @xmath108 ( dashed lines ) is proportional to the average spin fluctuations @xmath11 ( dotted lines ) .
this is an intriguing advantage of the high spin @xmath109 as opposed to a pure spin-3/2 system . only for certain values of @xmath6 resonances with triply occupied sites
occur , while for a wide regime of parameters , the state remains fluctuationless . due to the additional components , spin fluctuations increase to @xmath51 even in deep lattices for very long evolution times .
the particle - number fluctuations , as discussed in section [ sec_measuren ] , as well as the population of the new spin - components are quantities that are accessible experimentally . in combination
they are direct evidence for superexchange processes in the absence of direct tunneling . in conclusion
, @xmath1 is a suitable atomic species for the proposed experiments to simulate superexchange hamiltonians in shallow lattices .
we have found a mechanism for fermionic high - spin systems that allows for the investigation of superexchange processes in shallow lattices , i.e. , at large amplitudes @xmath4 .
this regime corresponds to much faster time scales than in deep lattices and was not accessible yet .
the non - equilibrium dynamics of an initially band - insulating state with two components has been calculated , and we found the intersite dynamics to be fully dominated by the superexchange .
although , in general , spin - changing collisions circumvent pauli blocking and therefore allow for strong particle - number fluctuations in shallow lattices , we have shown that the first - order tunneling is strongly suppressed in few - well potentials while strong spin fluctuations arise via superexchange processes .
the underlying mechanism is driven by an emerging tunneling - energy gap that reestablishes an interaction gap , which is usually only expected in deep lattices .
furthermore , we have investigated the crossover of superexchange regimes with weak and strong spin fluctuations in deep and shallow lattices , respectively .
we have discussed the experimentally relevant case of optical double - well and plaquette potentials .
we have simulated the time evolution for a pure spin-3/2 system as well as for realistic parameters of @xmath1 and have shown that in both cases it is governed by the superexchange hamiltonian .
the latter allows for a measurement of spin fluctuations via the population of certain spin components . the double - well potential and the four - well plaquette are both suited for an experimental realization .
even larger lattices of up to six sites in one spatial direction show the existence of a sufficiently large tunneling - energy gap @xmath110 and a subordinate interaction - energy blockade .
the necessary experimental tools , i.e. , the optical potentials for a double - well in one and two dimensions @xcite as well as the preparation of the initial band insulator with @xmath1 atoms have both been demonstrated successfully @xcite .
we would like to thank k. sengstock , l. mathey and j.s .
krauser for very fruitful discussions .
we thank the deutsche forschungsgesellschaft ( dpg ) for financial support within grant for 801 and grant grk 1355 . | we show that fermionic high - spin systems with spin - changing collisions allow to monitor superexchange processes in optical superlattices with large amplitudes and strong spin fluctuations . by investigating the non - equilibrium dynamics
, we find a superexchange dominated regime at weak interactions .
the underlying mechanism is driven by an emerging tunneling - energy gap in shallow few - well potentials . as a consequence , the interaction - energy gap that is expected to occur only for strong interactions in deep lattices
is reestablished . by tuning the optical lattice depth ,
a crossover between two regimes with negligible particle number fluctuations is found : first , the common regime with vanishing spin - fluctuations in deep lattices and , second , a novel regime with strong spin fluctuations in shallow lattices .
we discuss the possible experimental realization with ultracold @xmath0k atoms and observable quantities in double wells and two - dimensional plaquettes . |
following the work of fontana and srivastava @xcite , agarwal @xcite , cardimona , raymer , and stroud @xcite , and zhu and scully @xcite , a number of articles have appeared containing proposals for suppressing spontaneous emission @xcite .
in contrast to the suppression of spontaneous emission that one can achieve by placing an atom in a cavity whose radiation modes differ from those of free space , it is suggested in these articles that spontaneous emission in free space can be suppressed by applying an external radiation field to an atom having a specified level scheme .
this is a rather remarkable result , since one might imagine that , owing to the very short correlation time of the vacuum field , such modification of spontaneous emission rates would be strictly forbidden .
a prototypical level scheme that leads to suppression of spontaneous emission is that of zhu and scully @xcite ( see fig . [ fig1 ] ) .
two excited states @xmath0 and @xmath1 are separated in frequency by @xmath2 .
these states decay to the ground state @xmath3 with rates @xmath4 and @xmath5 , respectively .
what makes the decay scheme somewhat unusual is that states @xmath6 and @xmath1 are _ coupled directly by the vacuum field_. an external radiation field couples an auxiliary , _ metastable _
state @xmath7 to both states @xmath6 and @xmath8 for certain values of the field strength and atom - field detunings , it is found that one can have a nonvanishing , significant , steady state probability for the atom to be in states @xmath9 or @xmath1 . as such , spontaneous emission from these levels
is suppressed by the presence of the driving field .
_ @xcite claim to have observed this effect in an experiment on sodium dimers .
= 8 cm = 8 cm the suppression of spontaneous emission has been explained in terms of a dressed state of the atom - field system that is decoupled from the vacuum radiation field @xcite .
how is this decoupling accomplished ?
is there any underlying structure in the proposed level schemes that can help one to understand this most surprising result ?
it is the purpose of this article to address these questions . by considering a model problem
, i will show that states @xmath9 and @xmath10 can be viewed as superpositions of two states , _ one of which is metastable_. it is this metastable state that is necessary for the total suppression of spontaneous emission .
moreover , the decay dynamics required for total suppression of spontaneous emission implies that the level scheme of fig .
[ fig1 ] is isomorphic to a three - level lambda system .
the lower two levels of the lambda system are _ both _ metastable , and each is coupled to the decaying state . as such , the total suppression of spontaneous emission can be explained in terms of conventional dark states and coherent population trapping @xcite .
the experiment of xia et al @xcite will also be discussed .
while the level scheme they study is relevant to this class of problems , the results they obtained can not be classified as a suppression of spontaneous emission .
in the absence of the driving field , the equations for the evolution of the state amplitudes @xmath11 and @xmath12 given by zhu and scully @xcite are [ 1 ] @xmath13 where @xmath14 the first question we must ask is whether or not these equations correctly describe the interaction of an atom with the vacuum radiation field .
the answer to this question is not obvious . if we consider the energy levels shown in fig .
[ fig1 ] to be those of an isolated atom in free space , we immediately run into some problems . from the dipole selection rules , it is easy to show that the vacuum coupling from state @xmath9 to state @xmath1 must conserve orbital , spin - orbit , and total angular momenta @xmath15 , @xmath16 , and @xmath17 , as well as the z - component of total angular momentum . as a consequence , states @xmath6 and @xmath1 _ must belong to different electronic state manifolds_@xmath18 this , in turn , implies that @xmath2 corresponds to a frequency that is orders of magnitude larger than the decay rates @xmath4 and @xmath5 , respectively .
the rapid oscillation of state amplitudes @xmath11 and @xmath12 with frequency @xmath2 brings into question the validity of the weisskopf - wigner approximation used for the derivation of eqs .
. there is , perhaps , a more subtle point involved . starting from state @xmath9 , one can emit a photon taking the atom to state @xmath19 reabsorb this photon taking the atom to _ virtual state _
@xmath20 reemit a photon taking the atom to state @xmath21 and reabsorb this photon returning the atom to state @xmath9 .
this overall process constitutes an @xmath22rydberg ) contribution to the lamb shift of state @xmath9 .
consequently , if the atomic states are renormalized to include the lamb shift , it is questionable as to whether the vacuum coupling in between states @xmath9 and @xmath1 should be included in eqs .
( [ 1 ] ) @xcite .
it thus appears unlikely that one can achieve the vacuum coupling indicated in eqs .
( [ 1 ] ) if these states correspond to eigenstates of a free , isolated atom , dressed by the vacuum field . on the other hand
, it _ is _ possible to achieve this vacuum coupling if states @xmath9 and @xmath1 correspond to eigenstates of an atom plus some external field or , in some cases , to the states of a molecule @xcite .
the most obvious atom candidate is a hydrogen atom in a static electric field @xcite .
states @xmath9 and @xmath1 could then be chosen as linear combinations of the 2s and 3p states of hydrogen . the idea of using a hydrogen atom in a static electric field to modify the spontaneous emission spectrum is not new .
zhu and scully mention it in their 1996 article @xcite , and fontana and srivastava gave a detailed analysis of the decay in their 1973 article @xcite .
alternatively , one could use a level scheme similar to that used by xia _
@xcite , in which states @xmath9 and @xmath10 are superposition of singlet and triplet states in a molecule .
i will return to the experiment of xia _ et al .
iii . in order to gain additional insight into this problem
, i consider the level scheme shown in fig .
states @xmath21 , @xmath23 @xmath24 , @xmath25 , and @xmath26 are eigenstates of an unperturbed hamiltonian . states @xmath24 and @xmath25 have opposite parity and are coupled by a constant potential @xmath27 @xcite . an external radiation field , having frequency @xmath28 couples state @xmath29 , which is assumed to be metastable , to state @xmath30 only .
states @xmath24 and @xmath31 decay to states @xmath21 and @xmath32 with rates @xmath33 and @xmath34 , respectively .
( note that this level scheme could correspond to hydrogen with state @xmath24 corresponding to @xmath35 , state @xmath25 to @xmath36 , and states @xmath21 , @xmath37 and @xmath7 to @xmath38 @xcite . in this case , @xmath39 for the @xmath40 state@xmath18 ) the goal of this calculation is to show that the level scheme of fig .
[ fig2 ] , a level scheme exhibiting conventional decay dynamics , can be mapped into that of fig .
[ fig1 ] , a level scheme exhibiting somewhat unconventional decay dynamics .
thus , the suppression of spontaneous emission can equally well be analyzed using the level schemes of fig .
[ fig1 ] or [ fig2 ] .
it will be seen that the suppression of spontaneous emission can be explained in terms of conventional dark states when the level scheme of fig .
[ fig2 ] is used .
the mapping between the two level schemes is achieved by identifying states @xmath9 and @xmath1 , appearing in fig .
[ fig1 ] , as eigenstates of the unperturbed hamiltonian associated with fig .
[ fig2 ] , plus the potential @xmath41 = 8 cm = 8 cm in the absence of decay , the effective hamiltonian for the level scheme of fig .
[ fig2 ] , in the rotating wave approximation , in an interaction representation , and with the energy of level @xmath42 taken equal to zero , can be written as @xcite @xmath43 where the order of the states is ( @xmath7,@xmath30,@xmath44 , @xmath45 is a rabi frequency ( taken to be real ) and @xcite @xmath46 states @xmath21 and @xmath47 have not been included in ( [ 3 ] ) since they are not needed for the present discussion of the decay dynamics of the excited states .
the hamiltonian ( [ 3 ] ) can be diagonalized without much difficulty ; however , the desired comparison between the level schemes of figs . 1 and 2 is achieved by diagonalizing the @xmath48 subspace only .
the new eigenstates are given by [ 5 ] @xmath49 where @xmath50 in terms of these eigenstates , the transformed hamiltonian takes the form @xmath51 where the order of the states is ( @xmath7,@xmath6,@xmath52 and a constant energy @xmath53 has been subtracted from the energy of each state .
the equations of motion for the state amplitudes are [ 9 ] @xmath54 it is now a simple matter to include decay into these equations .
since spontaneous decay is governed by @xmath55 ; @xmath56 where @xmath57 and since @xmath58 ; @xmath59 , it follows that eqs .
( [ 9 ] ) , including decay , can be written as [ 11 ] @xmath60 where [ 12 ] @xmath61 this form of the equations is _ almost _ identical to that used in theories of suppression of spontaneous decay [ compare with eq .
( [ 1 ] ) ] based on the level scheme of fig .
[ fig1 ] . for the equations to be_identical , _ and for the level schemes of figs . 1 and 2 to be isomorphic
, one must require that @xmath62 it follows from eqs .
( [ 12 ] ) that the only way this equation can be satisfied is to have @xmath63 .
in other words , the form of the vacuum coupling given in eqs .
( [ 1 ] ) for the level scheme of fig .
[ fig1 ] in theories of total suppression of spontaneous
_ necessarily implies that state _
@xmath64 _ of the equivalent level scheme _ _ of fig .
[ fig2 ] _ _ must be metastable_. since both states @xmath25 and @xmath7 are metastable and do not undergo spontaneous emission in the isolated atom , it is reasonable to ask whether or not the level scheme of fig .
[ fig1 ] legitimately qualifies to be labeled as one in which spontaneous emission has been suppressed . in order to determine
if the driving field suppresses spontaneous emission , one must first establish that spontaneous emission of states @xmath9 and @xmath1 always occurs in the _ absence _ of the driving field . setting @xmath65 in eqs .
( [ 11 ] ) , one finds that the only steady state solution is @xmath66 any initial state population in states @xmath9 and @xmath10 decays if the driving field is absent .
this is easily understood in terms of the original @xmath24 , @xmath31 basis ; although state @xmath25 is metastable , it is coupled to the decaying state @xmath24 by the potential @xmath41 no matter how weak the coupling strength @xmath67 , any initial population in state @xmath25 eventually leaks out via state @xmath68 does the presence of the driving field suppress this spontaneous emission ?
the answer to this question is affirmative if the initial state is an arbitrary superposition of states @xmath9 and @xmath10 and their remains population trapped in states @xmath6 and @xmath1 as the time approaches infinity .
an initial condition in which the atom is in state @xmath69 corresponding to the initial condition in the experiment of xia _ et al . _
@xcite , can not be used directly to establish total suppression of spontaneous emission , but can provide indirect evidence for this effect , as discussed in sec .
iii below .
having established that state @xmath25 must be metastable to satisfy the requirements for spontaneous emission suppression , it is now an easy matter to understand the total suppression of spontaneous emission by returning to the original hamiltonian ( [ 3 ] ) .
an inspection of this hamiltonian reveals that it is identical to a hamiltonian , written in a field interaction representation @xcite , that characterizes a three - level atom in a lambda scheme driven by two fields .
the field having rabi frequency @xmath45 and detuning @xmath70 drives the @xmath71-@xmath42 transition and the field having rabi frequency @xmath67 and detuning @xmath72 drives the @xmath42-@xmath73 transition @xcite .
total suppression of spontaneous emission occurs if one can find an eigenstate consisting of a superposition of state amplitudes @xmath74 and @xmath75 which is decoupled from state amplitude @xmath76 .
this _ dark state _
@xcite does not decay since it is a superposition of nondecaying states .
in other words we seek values of @xmath77 and @xmath78 for which the superposition of state amplitudes of the form @xmath79 satisfies the equation of motion @xmath80 from schrdinger s equation with the hamiltonian ( [ 3 ] ) , it follow that @xmath81 equation ( [ 16 ] ) can be satisfied only if @xmath82 @xmath83 which implies that @xmath84 .
the driving field must be tuned to the frequency that would correspond to a hole in the emission spectrum from the 2 - 3 state manifold @xcite .
thus if @xmath85 there always exists a dark state amplitude of the system @xmath86 where @xmath87 which does not decay .
the other eigenstate amplitudes [ 14 ] @xmath88 where @xmath89 contain an admixture of state amplitude @xmath76 and decay as @xmath90 . as a consequence , any initial condition for which @xmath91 has a metastable component that does not decay as the time approaches infinity .
it remains only to establish that an initial condition of the form @xmath92 leads to a final state which has some population trapped in states @xmath9 and @xmath1 ( or , equivalently , in state @xmath93 as @xmath90 , the solution for the total dressed state amplitudes @xmath94 is @xmath95 @xmath96 @xmath97 , which , when reexpressed in terms of the bare state initial conditions [ with @xmath98 $ ] is @xmath99 . the final state populations are [ 21 ] @xmath100 where eqs .
( [ 4 ] ) , ( [ 5p ] ) and ( [ 19s ] ) were used .
thus , we see that population is always trapped in states @xmath9 and @xmath101 equations ( [ 21 ] ) for the probabilities @xmath102 , @xmath103 , and @xmath104 can be written in terms of the couplings and detunings in the ( @xmath29,@xmath9,@xmath52 basis . referring to eq .
( [ 7 ] ) and using eqs .
( [ 4 ] ) , ( [ 5p ] ) , ( [ 12d ] ) , ( [ 18 ] ) , ( [ 19s ] ) , one finds the appropriate relationships , @xmath105 @xmath106 @xmath107 , and @xmath108 @xmath109 @xmath110 , subject to the constraint , @xmath111 . under conditions of total suppression of spontaneous emission , the field
is tuned to the energy of the metastable level @xmath73 ; that is , @xmath112 and @xmath113 . as an aside
, i might note that the results can be reinterpreted as a suppression of absorption @xcite if one starts with all population initially in state @xmath7 , for which @xmath114 as @xmath90 . in the absence of the coupling potential @xmath27 ,
the steady state population @xmath115 would vanish , but it does not vanish in the presence of this coupling . in this sense , it is closely related to electromagnetically induced transparency @xcite .
it has been shown that the origin of the suppression of spontaneous emission proposed by zhu and scully @xcite and others @xcite can be traced to a metastable state that is hidden in their calculations .
once this hidden state is revealed , the suppression of spontaneous emission can be understood in terms of a conventional dark state and coherent population trapping @xcite that can arise when an atom having a three - level , lambda scheme is driven by two fields . the dark state in this instance is a superposition of two metastable states so is , itself , metastable .
the dynamical suppression of spontaneous emission is a real effect .
if the external driving field @xmath45 were not present , the two state manifold consisting of states @xmath9 and @xmath1 would always decay . in some sense , the driving field allows one to access the metastable level @xmath25 contained in both states @xmath9 and @xmath116 .
this type of dynamical suppression could be used , for example , to reduce spontaneous emission in the @xmath40-@xmath117 manifold of hydrogen resulting from stray fields that couple the @xmath118 state to the @xmath119 state .
it would be necessary to drive the @xmath40-@xmath117 transition using an _ rf _ field and the @xmath120-@xmath117 with a uv field having frequency @xmath121 @xcite .
the use of the equivalent ( @xmath122 basis rather than the ( @xmath123 basis greatly simplifies the interpretation of the results . from the analysis of sec .
ii , it is clear that the final state probabilities depend only on the initial state amplitude @xmath124 /r_{b}$ ] and not on the decay rate if state @xmath24 decays to state @xmath3 only .
if state @xmath24 decays to state @xmath26 as well as to state @xmath21 , or if states @xmath7 and @xmath21 actually correspond to the same state@xmath125 the final state probabilities are modified , but the steady state still corresponds to a dark state for which there is total suppression of absorption . on the other hand , if state @xmath29 is not metastable , there can not be total suppression of spontaneous emission since the dark state amplitude @xmath126 /r_{b}$ ] decays to zero as @xmath90 .
= 8 cm = 8 cm finally i should like to discuss the experiment of xia _ et al . _
@xcite _ , _ who used the level scheme shown in fig .
[ fig3 ] , corresponding to molecular states in the sodium dimer .
states @xmath9 and @xmath127 are superpositions of singlet and triplet states which are mixed by a spin - orbit interaction . in the spirit of this calculation , one can associate the singlet and triplet states with states @xmath128 and @xmath129 respectively , in the hamiltonian ( [ 3 ] ) , and the spin - orbit mixing with the potential @xmath27 . of course , it is not possible to turn off the mixing potential in this case .
the singlet component of states @xmath9 and @xmath10 decays to state @xmath21 and the triplet component of states @xmath9 and @xmath1 decays to state @xmath47 , while the singlet component of states @xmath9 and @xmath1 is driven by a two - photon transition from the ground state . since _ both _ the singlet and triplet components decay ,
the conditions for suppression of spontaneous emission are not met ( recall that it was necessary that state @xmath130 which corresponds to the triplet state , be metastable ) . in their experiment , xia _ et al .
_ are not measuring spontaneous _ _ emission , as it is normally defined .
instead , they are measuring _ scattering _ via the _ three - photon _ process in which two photons are absorbed from the driving field and a vacuum photon
is emitted taking the atom to state @xmath21 ( singlet channel ) or @xmath32 ( triplet channel ) .
they found that , for a tuning of the incident field midway between levels @xmath9 and @xmath127 , 2@xmath131 , the scattering in the singlet channel was suppressed and that in the triplet channel was enhanced .
this constitutes strong evidence that states @xmath0 and @xmath1 are coupled directly by the vacuum field , and that the singlet and triplet states are degenerate in the absence of the spin - orbit coupling ( @xmath132 @xmath133 ) @xcite .
although this experiment is important insofar as it provides an example of a system in which vacuum coupling of two , distinct excited states occurs , it does not demonstrate suppression of spontaneous emission
. there will be no steady - state population in states @xmath9 and @xmath127 .
on the other hand , xia _ et al .
_ have shown that scattering in a specific channel can be suppressed . as was noted above
, total suppression of absorption occurs under the same conditions as total suppression of spontaneous emission , so that the existence of one implies the other .
consequently , if one can demonstrate total suppression of absorption , the system will also exhibit total suppression of spontaneous emission . to establish total suppression of absorption
, one can either ( i ) prove that there is no absorption of the driving field or ( ii ) show that there is no scattered radiation for _ all _ polarizations and directions of the scattered field .
the existence of scattered radiation in the triplet channel in the experiment of xia _ et al .
_ necessarily implies that there is _ not _ total suppression of absorption . on the other hand
, the absorption rate from the ground state is decreased by a factor @xmath134/(@xmath135 ) relative to that which would have occurred if states @xmath9 and @xmath127 were not coupled by the vacuum field .
consequently , one can say that the spontaneous emission rate or the absorption rate is decreased in this system if @xmath136triplet)@xmath137singlet ) .
the data seems to indicate that @xmath138 and @xmath139 are comparable .
i am pleased to acknowledge helpful discussions with j. l. cohen , b. dubetsky , p. milonni , and g. w. ford .
this work is supported by the national science foundation under grant no .
phy-9414020 and the u. s. office of army research under grant no . daag55 - 97 - 0113 . for a discussion of this topic , see d. a. cardimona and c. r. stroud , jr .
a * 27 * , 2456 ( 1983 ) .
cardimona and stroud suggest that appropriate level schemes may also be found in excited states of high z , hydrogenic ions and in multielectron atoms , but give no specific examples .
if the levels @xmath42 and @xmath73 are separated by a radio frequency , one can replace the static field by a radio frequency field . in that case one
sets @xmath140 and replaces @xmath67 by the rabi frequency of the _ rf _ field ( assuming the rabi frequency is much less than @xmath141 ) . normally , the reduction of scattering in a _
single _ direction does not constitute evidence for the vacuum coupling of the excited states .
the scattering must vanish in _ all _ directions .
for example , if an incident field is tuned between the 3p@xmath143 and 3p@xmath144 states of na , there can be a total suppression of scattering for specific directions and polarizations of the scattered field , even though there is no vacuum coupling of these states .
the decrease in scattering in some directions is compensated by an increase in other directions . in the experiment of xia et al .
, however , the existence of scattering in the triplet channel offers proof that states @xmath6 and @xmath1 each contain an admixture of singlet and triplet states ; as such , each can be coupled to states @xmath3 ( or @xmath47 ) , leading to a vacuum coupling of states @xmath9 and @xmath145 | it has been shown recently [ see , for example , s .- y .
zhu and m. o. scully , phys .
rev .
lett . * 76 * , 388 ( 1996 ) ] that a dynamical suppression of spontaneous emission can occur in a three - level system when an external field drives transitions between a metastable state and _ two _ decaying states .
what is unusual in the decay scheme is that the decaying states are coupled directly by the vacuum radiation field .
it is shown that decay dynamics required for total suppression of spontaneous emission necessarily implies that the level scheme is isomorphic to a three - level lambda system , in which the lower two levels are _ both _ metastable , and each is coupled to the decaying state . as such , the total suppression of spontaneous emission can be explained in terms of conventional dark states and coherent population trapping . 0.4 in epsf |
one of the most popular approaches to constraining the mass density parameter @xmath0 , the ratio of the average matter density to the critical density , is based on comparisons between the galaxy density field mapped by redshift surveys and the galaxy peculiar velocity field inferred from distance - indicator surveys ( see the review by @xcite ) . while the numerous implementations of this approach differ in many details , they are all motivated by the linear theory formula for the peculiar velocity field , @xmath3 or its divergence @xmath4 where @xmath5 is the mass density contrast , @xmath6
, @xmath7 is the hubble parameter , and @xmath8 is the present value of the expansion factor ( @xcite ) . and @xmath8 do not introduce any uncertainy in peculiar velocity predictions ; if one adopts km s@xmath9 distance units in place of mpc , then @xmath7 and @xmath8 do not appear in equation ( [ eqn : vlin ] ) or ( [ eqn : divv ] ) .
] `` velocity - velocity '' comparisons start from the observed galaxy density field , predict peculiar velocities via equation ( [ eqn : vlin ] ) or some non - linear generalization of it , and compare to estimated peculiar velocities ( e.g. , @xcite ; @xcite ; @xcite ; @xcite , 1998 ; @xcite ) .
`` density - density '' comparisons start from the observed radial peculiar velocity field , infer the 3-dimensional velocity field using the potent method of bertschinger & dekel ( 1989 ) , and compare the velocity divergence to the observed galaxy density field using equation ( [ eqn : divv ] ) or a non - linear generalization of it ( e.g. , @xcite ; @xcite ; @xcite ; @xcite ) . because the radial velocity field must be smoothed before computing the 3-dimensional velocity field via potent , density - density comparisons in practice always compare the smoothed galaxy density field to predictions derived from the smoothed peculiar velocity field .
velocity - velocity comparisons , on the other hand , usually smooth the galaxy density field to suppress non - linear effects and shot noise , but compare the velocity predictions from these smoothed density fields directly to the estimated peculiar velocities of individual galaxies or groups . ( the spherical harmonic analysis of davis et al .
1996 is an important exception in this regard . )
the avoidance of smoothing the data is often seen as an advantage of the velocity - velocity approach , since smoothing a noisy estimated velocity field can introduce statistical biases that are difficult to remove .
however , in this paper we show that comparing smoothed velocity predictions to unsmoothed velocity measurements generally leads to biased estimates of @xmath10 , even when the galaxy positions and velocities are known perfectly .
the reason for this bias is fairly simple : the errors in the predicted velocities are correlated with the predicted velocities themselves , violating the conventional assumption that an individual galaxy s velocity can be modeled as a `` large scale '' contribution predicted from the smoothed density field plus an _ uncorrelated _ `` small scale '' contribution .
galaxy redshift surveys map the galaxy density field @xmath11 rather than the mass density field @xmath12 , so inferences from velocity - velocity and density - density comparisons often assume a linear relation between galaxy and mass density contrasts , @xmath13 , and therefore constrain the quantity @xmath14 rather than @xmath10 itself .
the results reported in this paper emerged from a more general investigation of the effects of complex galaxy formation models on estimates of @xmath15 ( @xcite ) . however , the statistical bias in @xmath10 that we find applies even when galaxies trace mass exactly , so here we focus on this simpler case .
( throughout this paper we use the term `` bias '' to refer to systematic statistical errors rather than the relation between the distributions of galaxies and mass . )
we further restrict our investigation to the case in which galaxy positions and velocities are known perfectly , ignoring the additional complications that arise in analyses of observational data .
we have carried out n - body simulations of three different cosmological models , all based on inflation and cold dark matter ( cdm ) .
the first is an @xmath16 , @xmath17 model ( @xmath18 ) , with a tilted power spectrum of density fluctuations designed to satisfy both cobe and cluster normalization constraints .
the cluster constraint requires @xmath19 ( @xcite ) , where @xmath20 is the rms linear density fluctuation in spheres of radius @xmath21mpc .
matching the cobe - dmr constraint and @xmath22 with @xmath17 requires an inflationary spectral index @xmath23 if one incorporates the standard inflationary prediction for gravitational wave contributions to the cobe anisotropies ( see @xcite and references therein ) .
the other two models have @xmath24 and @xmath25 , with a power spectrum shape parameter @xmath26 ( in the parameterization of @xcite ) and cluster - normalized fluctuation amplitude @xmath27 .
we ran four independent simulations for each of the three cosmological models , and the results we show below are averaged over these four simulations .
all simulations were run with a particle - mesh ( pm ) n - body code written by c. park , which is described and tested by @xcite .
each simulation uses a @xmath28 force mesh to follow the gravitational evolution of @xmath29 particles in a periodic cube @xmath30mpc on a side , starting at @xmath31 and advancing to @xmath32 in 46 steps of equal expansion factor @xmath33 .
we form the mass density field by cloud - in - cell ( cic ) binning the evolved mass distribution onto a @xmath29 grid .
we smooth this density field with a gaussian filter of radius @xmath34 and derive the linear - theory predicted velocity field using equation ( [ eqn : vlin ] ) . finally , we linearly interpolate this velocity field to the galaxy positions to derive predicted galaxy peculiar velocities @xmath35 .
figure 1 compares the true velocities of particles ( @xmath36 ) from one of the @xmath16 simulations to the velocities predicted ( @xmath35 ) by equation ( [ eqn : vlin ] ) from the mass density field smoothed with gaussian filters of radius @xmath37 and @xmath38 ( panels a - d , respectively ) .
the points in figure 1 show one cartesian component of the particles velocities .
if we make the assumption , common to most velocity - velocity comparison schemes , that each galaxy s velocity consists of a large scale contribution predicted from the density field plus an uncorrelated small scale contribution , then the best - fit slope of the @xmath39 relation should yield the parameter @xmath10 , in this case @xmath40 , with the scatter about this line yielding the dispersion of the small scale contribution . however , it is clear from figure 1 that this slope increases systematically with increasing @xmath34 .
( we note that the best - fit line , which minimizes @xmath41 , is shallower than the line one would naively draw through these data points by eye , since it is vertical scatter rather than perpendicular scatter that must be minimized . )
the filled points in figure 2 show the estimated @xmath10 as a function of @xmath34 for the @xmath16 ( circles ) and @xmath24 ( squares ) cosmological models .
the solid lines show the true value of @xmath10 . in both cases ,
the estimated value of @xmath10 is quite sensitive to the smoothing scale : it is slightly underestimated at small scales , but increasingly overestimated at large scales .
the @xmath42 model yields similar results , so we do not plot it separately .
we also investigated @xmath16 simulations with a factor of two lower force resolution ( @xmath29 force mesh instead of @xmath28 ) and found identical results , so even at small smoothing scales our results are not affected by the simulations limited gravitational resolution .
the breakdown of linear theory at small scales is not surprising ; however , the systematic failure of this method at large smoothing scales has not , to our knowledge , been previously discussed .
the dependence of the estimated @xmath10 on the smoothing scale used for velocity predictions is our principal result .
we can understand the origin of the large scale bias in @xmath10 by considering the case in which galaxy peculiar velocities are given exactly by linear theory .
in this case , @xmath43 where @xmath44 are the fourier modes of the density field and the integral extends over all of @xmath45-space . predicted velocities , however , are estimated from the density field smoothed with a window function @xmath46 of characteristic scale @xmath34 .
therefore , @xmath47 where @xmath48 is the fourier transform of the window function .
the error in the predicted velocity of a galaxy at position @xmath49 is therefore , @xmath50
e^{i { { \bf k}}\cdot { { \bf x } } } \frac{i\delta_{{{\bf k } } } { { \bf k}}}{|{{\bf k}}|^2 } d{{\bf k}}~. \label{eqn : dv}\ ] ] note that in equation ( [ eqn : vpred ] ) we have defined @xmath35 to be the velocity that would be predicted assuming the correct value of @xmath0 . in practice , since we do not know the value of @xmath10 beforehand , we derive its value from the slope of the @xmath36 vs. @xmath51 relation ( this is equivalent to assuming @xmath16 when computing @xmath35 ) . if @xmath52 were uncorrelated with @xmath35 , then the slope of the @xmath36 vs. @xmath51 relation would be an unbiased estimator of @xmath10
. however , if @xmath52 is positively correlated with @xmath35 , then the slope of the relation is no longer @xmath10 , since points preferentially scatter above the line for positive @xmath35 and below the line for negative @xmath35 .
this steepening of the @xmath39 relation is just the behavior seen in figure 1 .
equations ( [ eqn : vpred ] ) and ( [ eqn : dv ] ) show that @xmath52 and @xmath35 will be correlated as long as some fourier modes contribute to both integrals , which happens for any smoothing function other than a step function in @xmath45-space
. we can quantitatively understand this bias by considering how @xmath10 is measured . for an ensemble of @xmath53 points ( @xmath54 , @xmath55 ) , the slope of the best - fit line ( assuming @xmath56 ) is @xmath57 } { \frac{1}{n } \sum ( { { \bf v}_{\mathrm{pred},\mathit{i}}}\cdot { { \bf v}_{\mathrm{pred},\mathit{i } } } ) } \nonumber \\ & = & f(\omega)\left[1 + \frac{\langle\delta
{ { \bf v}}\cdot { { \bf v}_{\mathrm{pred}}}\rangle } { \langle{{\bf v}_{\mathrm{pred}}}\cdot { { \bf v}_{\mathrm{pred}}}\rangle}\right ] ~. \label{eqn : slope}\end{aligned}\ ] ] equation ( [ eqn : slope ] ) shows how a non - zero cross - correlation between @xmath52 and @xmath35 changes the measured slope of the velocity - velocity relation .
we can compute this effect in the linear regime for a given power spectrum of density fluctuations @xmath58 and window function @xmath48 . using equations ( [ eqn : vpred ] ) and ( [ eqn : dv ] ) we have @xmath59 p(k )
dk } { \int_0^\infty \widetilde{w}^2(kr_s ) p(k ) dk } ~. \label{eqn : dvvpred}\end{aligned}\ ] ] for gaussian and top hat window functions and a range of cdm power spectra , we find that the bias given by equation ( [ eqn : dvvpred ] ) is always positive and is always an increasing function of @xmath34 .
the dashed lines in figure 2 show the slope computed ( from eqs .
[ eqn : slope ] and [ eqn : dvvpred ] ) using the linear mass power spectra of the simulations and the same gaussian window functions that were used to measure @xmath10 ( solid points ) .
the striking similarity on large smoothing scales between the n - body data and this linear theory calculation supports our conclusion that the large scale bias is indeed caused by the cross - correlation between @xmath52 and @xmath35 , which , in turn , is caused by the comparison of a smoothed prediction to unsmoothed data . from equation ( [ eqn : dvvpred ] ) it is evident
that the linear theory cross - correlation between @xmath52 and @xmath35 will be equal to zero if there is no smoothing at all , or if the smoothing function is a step function in @xmath45-space , in which case the product @xmath60 $ ] is always equal to zero .
the open symbols in figure 2 show results of a velocity - velocity analysis of the same simulations , with the linear theory velocities now predicted from a density field smoothed with a sharp , low - pass @xmath45-space filter . specifically , we set to zero all fourier modes with @xmath61 and plot the new estimates of @xmath10 at the values of @xmath34 for which a gaussian filter falls to half its peak value at @xmath62 ( i.e. , @xmath63 ) . using the sharp @xmath45-space filter causes the bias to vanish completely on large scales , yielding estimates of @xmath10 that are correct and independent of smoothing length .
this result further supports our interpretation of the cause of the large scale velocity - velocity bias .
figure 2 shows that @xmath10 is underestimated at small scales in the n - body simulations .
the linear theory bias discussed above and shown by the dashed line in figure 2 is always positive .
therefore , there must be a countervailing effect that biases @xmath10 estimates in the opposite direction on small scales . in highly non - linear regions of the density field , such as the cores of galaxy clusters , linear theory velocity predictions have large errors .
however , errors caused by virial motions are uncorrelated with the predicted velocities because these virial motions have random directions .
such errors add random scatter to the velocity - velocity relation , but they do not change its slope . in mildly non - linear regions of the density field ,
on the other hand , galaxy velocities still follow coherent flows , but these flows may no longer be accurately predicted by linear theory . in the case of a galaxy falling towards a large over - density , linear theory will correctly predict the direction of motion , but it will overestimate the infall speed because it incorrectly assumes that the over - density has grown at the linear theory rate over the history of the universe , while in reality the over - density grows to large amplitude only at late times when it becomes non - linear . in such regions , @xmath52 will be opposite in sign to @xmath35 , causing an anti - correlation between the two quantities .
the opposite happens in under - dense regions , but since fewer galaxies reside in these regions and the velocity errors are smaller in magnitude , the net effect is still an anti - correlation between @xmath52 and @xmath35 . in order to show how these different effects come into play , we adopt a fluid dynamics description and divide an individual galaxy s velocity into a mean flow @xmath64 and a random `` thermal '' velocity @xmath65 , so that @xmath66 . here
@xmath67 is the average velocity of galaxies at spatial position @xmath49 , and therefore @xmath68 0 by definition .
let @xmath69 denote the velocity predicted in linear theory from the _ unsmoothed _ density field ( eq .
[ eqn : vlin ] ) .
equation ( [ eqn : dv ] ) applies to the case where the velocity field is exactly linear , @xmath70 , but more generally , @xmath71 and , therefore , @xmath72 this equation shows the three possible sources of error in the smoothed linear theory prediction of galaxy velocities .
the first term represents the effect caused by comparing a smoothed quantity with an unsmoothed quantity in linear theory and is given by equation ( [ eqn : dv ] ) .
the second term represents the inadequacy of using a linear theory velocity estimator in regions where non - linear effects are important .
the third term represents errors caused by galaxies random thermal motions . as shown in equation ( [ eqn : slope ] ) , the bias in @xmath10 depends on the cross - correlation of these errors with @xmath35 , @xmath73 the first term is positive and causes an overestimate of @xmath10 for nearly all smoothing functions .
our calculation of this effect via equation ( [ eqn : dvvpred ] ) shows that it is zero for no smoothing and increases monotonically with smoothing scale .
we have argued above that the second term is generally negative and causes an underestimate of @xmath10 .
since this effect arises from the non - linearity in the density field , it should dominate on small scales and vanish with increased smoothing of the density field . finally , the third term is equal to zero because the thermal velocities have random directions .
a combination of the first two terms of equation ( [ eqn : dvv3 ] ) explains the scale dependence of @xmath10 estimates in figure 2 . for large smoothing of the density field
, the first term dominates and we overestimate @xmath10 , whereas for small smoothing the second term dominates and we underestimate @xmath10 .
the estimate of @xmath10 is unbiased at the smoothing scale where these two effects cancel , but this scale should itself depend on the specifics of the underlying cosmological model . the numerical results in figure 2 confirm this prediction : the @xmath10 estimate is unbiased at @xmath74 in the @xmath24 model ( with @xmath75 ; squares ) and at @xmath76 in the @xmath16 model ( with @xmath22 ; circles ) .
the smoothing scale for unbiased estimates could also depend on the assumed relation between galaxies and mass , a point we will investigate in future work .
it is therefore , not possible to remove this bias simply by choosing the right smoothing scale in a model - independent way . if we had adopted a higher - order perturbative expansion for predicting velocities from the smoothed density field , then equation ( [ eqn : dvv3 ] ) would still hold with @xmath69 replaced by @xmath77 , the perturbative prediction in the absence of smoothing .
the first term on the right hand side would still be positive , since some fourier modes would contribute to both ( @xmath78 ) and @xmath35 . the second term could be positive or negative depending on the approximation and the smoothing scale .
however , while a higher - order approximation might reduce the magnitude of the second term relative to the linear approximation , it would not necessarily reduce the net bias in @xmath10 , since this depends on the relative magnitude and sign of the first two terms .
the implications of our results for existing estimates of @xmath10 ( or , more generally , of @xmath15 ) are probably limited .
as already mentioned , density - density comparisons via potent are not influenced by the effects discussed here , because they compare density and velocity divergence fields smoothed at the same scale .
the analysis of davis et al .
( 1996 ) , a mode - by - mode comparison of density and velocity fields , is also not affected , since the two fields are again compared at the same effective `` smoothing '' . if the observed velocities are unsmoothed , a comparison in which velocities are predicted using a truncated spherical harmonic expansion of the density field ( e.g. , @xcite ) may behave rather like our sharp @xmath45-space filter analysis ( open symbols in figure 2 ) , since for a gaussian field the different spherical harmonic components are statistically uncorrelated ( a. nusser , private communication ; @xcite ) . among recent velocity - velocity studies ,
our procedure here is closest to the velmod analyses of willick et al .
( 1997 ) and willick & strauss ( 1998 ) , who used a @xmath79 gaussian filter to compute the predicted velocity field .
these authors chose their smoothing scale partly on the basis of tests on n - body mock catalogs , and our results in figure 2 suggest that biases in @xmath10 should indeed be small for this smoothing .
however , we have shown that the disappearance of the bias in @xmath10 at this smoothing scale occurs because of a cancellation between positive and negative biases , and that the scale at which this cancellation occurs depends at least to some degree on the underlying cosmological model . as improvements in observational data reduce the statistical uncertainties in peculiar velocity data , control of the systematic uncertainties that arise from comparing smoothed velocity predictions to unsmoothed data will become essential to obtaining robust estimates of the density parameter .
we thank adi nusser and michael strauss for helpful input and comments and marc davis and jeff willick for comments on the draft manuscript .
this work was supported by nsf grant ast-9802568 .
vkn acknowledges support by the presidential fellowship from the graduate school of the ohio state university .
berlind , a. a. , narayanan , v. k. , & weinberg , d. h. 1999 , in asp conference series , cosmic flows , eds .
s. courteau , m. a. strauss & j. willick ( asp : san francisco ) , astro - ph/9908208 bertschinger , e. , & dekel , a. 1989 , , 336 , l5 blakeslee , j. p. , davis , m. , tonry , j. l. , dressler , a. , & ajhar , e. a. 1999 , , in press , astro - ph/9910340 cole , s. , weinberg , d. h. , frenk , c. s. , & ratra , b. 1997 , , 289 , 37 davis , m. , nusser , a. , & willick , j. 1996 , , 473 , 22 dekel , a. , bertschinger , e. , yahil , a. , strauss , m. a. , davis , m. , & huchra , j. p. 1993
, , 412 , 1 dekel , a. , eldar , a. , kolatt , t. , yahil , a. , willick , j. a. , faber , s. m. , courteau , s. , & burstein , d. 1999 , , 522 , 1 efstathiou , g. , bond , j. r. , & white , s. d. m. 1992 , , 258 , 1p fisher , k. b. , lahav , o. , hoffman , y. , lynden - bell , d. , & zaroubi , s. 1995 , , 272 , 885 hudson , m. j. , dekel , a. , courteau , s. , faber , s. m. , & willick , j. a. 1995 , , 274 , 305 kaiser , n. , efstathiou , g. , saunders , w. , ellis , r. , frenk , c. , lawrence , a. , & rowan - robinson , m. 1991 , , 252 , 1 park , c. 1990 , phd thesis , princeton university peebles , p. j. e. 1980 , the large - scale structure of the universe , ( princeton : princeton university press ) sigad , y. , eldar , a. , dekel , a. , strauss , m.a .
, & yahil , a. 1998 , , 495 , 516 strauss , m. a. & willick , j. a. 1995 , physics reports , 261 , 271 | we show that a regression of unsmoothed peculiar velocity measurements against peculiar velocities predicted from a smoothed galaxy density field leads to a biased estimate of the cosmological density parameter @xmath0 , even when galaxies trace the underlying mass distribution and galaxy positions and velocities are known perfectly .
the bias arises because the errors in the predicted velocities are correlated with the predicted velocities themselves .
we investigate this bias using cosmological n - body simulations and analytic arguments . in linear perturbation theory , for cold dark matter power spectra and gaussian or top hat smoothing filters
, the bias in @xmath0 is always positive , and its magnitude increases with increasing smoothing scale .
this linear calculation reproduces the n - body results for gaussian smoothing radii @xmath1 , while non - linear effects lower the bias on smaller smoothing scales , and for @xmath2 @xmath0 is underestimated rather than overestimated .
the net bias in @xmath0 for a given smoothing filter depends on the underlying cosmological model .
the effect on current estimates of @xmath0 from velocity - velocity comparisons is probably small relative to other uncertainties , but taking full advantage of the statistical precision of future peculiar velocity data sets will require either equal smoothing of the predicted and measured velocity fields or careful accounting for the biases discussed here . |
a closed subgroup @xmath0 of the affine algebraic group @xmath1 is called an _ observable subgroup _ if the homogeneous space @xmath2 is a quasi - affine variety .
such subgroups have been researched extensively , notably by f. grosshans , see @xcite for a survey on this topic , and theorem [ thm : obsercond ] below for other useful characterizations of observable subgroups . in @xcite the authors presented the notion of an _ observable action _ of @xmath1 on the affine variety @xmath8 , together to its basic properties . in this paper
we develop further the notion of observable actions .
in particular , we investigate the situation where @xmath3 is an affine _ algebraic monoid _ with unit group @xmath1 , and @xmath0 is a closed subgroup of @xmath1 , such that the action of @xmath0 on @xmath3 by left multiplication is observable . in this case
, we say that @xmath0 is _ observable in @xmath3_. we describe now the organization of this paper .
in section [ sec2 ] we provide the basic definitions and properties of observable actions and affinized quotients . in section [ sec3 ]
we give several characterizations of observable subgroups . in theorem
[ thm : istabcond ] we deduce a number of useful consequences from the assumption that @xmath0 is observable in @xmath3 .
in particular it follows that @xmath0 is an observable subgroup of @xmath1 , and that it is closed in @xmath3 . in theorem [ theo:1 ]
we characterize the observable subgroups of @xmath3 in terms of semiinvariants . in theorem
[ thm : obserrep ] we show that @xmath0 is an observable subgroup of @xmath3 if and only if @xmath0 is the isotropy group of some vector @xmath9 in some rational representation @xmath10 of @xmath3 . in the final section ( section [ sec4 ] )
we use the results of the previous section to study the affinized quotient of an affine algebraic monoid by a closed normal subgroup . in theorem
[ thm : normalimplisobse ] we show that if @xmath0 is a closed , normal subgroup of @xmath1 , closed in @xmath3 , then @xmath0 is observable in @xmath3 .
whether this is true for nonnormal closed subgroups of @xmath1 is an open question .
see remark [ rem : openquest ] . as an application
, we show that the _ affinized quotient _ of an affine algebraic monoid @xmath3 by a normal subgroup @xmath0 , closed in @xmath3 , is an algebraic monoid , with unit group @xmath2 .
acknowledgements : this paper was written during a stay of the second author at the university of western ontario .
he would like to thank them for the kind hospitality he received during his stay .
let @xmath11 be an algebraically closed field .
we work with affine algebraic varieties @xmath8 over @xmath11 .
an algebraic group is assumed to be a smooth , affine , group scheme of finite type over @xmath11 . if @xmath8 is an affine variety over @xmath11 we denote by @xmath12 $ ] the ring of regular functions on @xmath8 . if @xmath13 $ ] is an ideal , we denote by @xmath14 .
if @xmath15 is a subset , we denote by @xmath16\mathrel { : } f(y)=0\ \forall\ , y\in y\bigr\}$ ] .
morphisms @xmath17 between affine varieties correspond to morphisms of algebras @xmath18 \to \bbbk[x]$ ] , by @xmath19 , @xmath20 . if @xmath8 is irreducible we denote by @xmath21 the field of rational functions on @xmath8 . if @xmath22 is any integral domain we denote by @xmath23 $ ] its quotient field . thus if @xmath8 is an irreducible affine variety , then @xmath24\bigr]$ ] .
let @xmath1 be an affine algebraic group and let @xmath8 be an algebraic variety .
a _ ( regular ) action _ of @xmath1 on @xmath8 is a morphism @xmath25 , denoted by @xmath26 , such that @xmath27 and @xmath28 for all @xmath29 and @xmath30 .
since all the actions we work with are regular , we drop the adjective regular .
the _ orbit _ of @xmath30 is denoted by @xmath31 . if @xmath32 is a regular left action of @xmath1 on @xmath8 we consider the induced right action of @xmath1 on @xmath12 $ ] , defined as follows . if @xmath33 $ ] and @xmath34 , then @xmath35 .
it is well known that @xmath1-stable closed subset of @xmath8 correspond to @xmath1-stable radical ideals of @xmath36 $ ] .
we say that @xmath37 $ ] is _ g - invariant _ if @xmath38 for any @xmath39 .
the set of @xmath1-invariants @xmath40 $ ] forms a @xmath11-subalgebra of @xmath12 $ ] , possibly non - finitely generated .
analogous considerations can be made if we start with a right action @xmath41 .
a _ finite dimensional ( rational ) @xmath1-module _ is a finite dimension @xmath42-vector space @xmath43 together with a left action of @xmath1 on @xmath43 by linear automorphisms . a right action of @xmath1 on @xmath43 defines , in a similar way ,
the notion of a _
right @xmath1-module_. recall that an _ algebraic monoid _
@xmath3 is an algebraic variety together with an associative product @xmath44 with neutral element @xmath45 , such that @xmath46 is a morphism of algebraic varieties .
we denote the _ set of idempotent elements _ of @xmath3 by @xmath47 .
we denote the _ unit group _ of @xmath3 by @xmath48 .
it is known that @xmath49 is an algebraic group , open in @xmath3 ( see @xcite and @xcite ) . in this section
we establish the basic facts about _ extendible characters _ for the case of linear algebraic monoids .
let @xmath3 be an algebraic monoid .
a _ character _ of @xmath3 is a morphism of algebraic monoids @xmath50 .
we denote the monoid of characters of @xmath3 by @xmath51\mathrel { : } \chi(ab)=\chi(a)\chi(b)\ \forall\ , a , b\in m\,,\ \chi(1)=1\bigr\}.\ ] ] if @xmath52 , then restriction induces an injective morphism of ( abstract ) monoids @xmath53 .
if @xmath3 is an irreducible affine algebraic monoid , then there exists @xmath54 and a morphism of algebraic monoids @xmath55 , such that @xmath56 is closed immersion ( see for example ( * ? ? ? * theorem 3.8 ) ) .
this motivates the following definition .
let @xmath3 be an irreducible affine algebraic monoid .
a character @xmath57 is called a _ determinant _ if @xmath58 . by the considerations above
, determinants always exists .
\(1 ) observe that if @xmath59 is a determinant , then @xmath60 .
in particular @xmath61=\bbbk[m]_{\det}$ ] .
\(2 ) clearly , @xmath62 is a unital submonoid that generates @xmath63 as a group .
indeed , let @xmath64 be a determinant and @xmath65 .
then there exists @xmath66 $ ] and @xmath54 such that @xmath67 .
it follows that @xmath68 is a character , and hence , by continuity , @xmath69 .
the notion of an _ extendible character _
is very useful in the study of observable subgroups of algebraic groups .
we extend this notion to the setting of algebraic monoids .
let @xmath3 be an affine algebraic monoid and let @xmath70 be a closed subgroup .
a non - trivial character @xmath71 is _ extendible _ if there exists a non - zero _ semiinvariant element of weight @xmath72 _ ; that is , there exists @xmath73 $ ] such that @xmath74 for all @xmath75 .
such an element is called an _ extension of @xmath72_. we denote the monoid of extendible characters of @xmath0 by @xmath76 .
clearly , restriction to @xmath52 induces an injective homomorphisms of ( abstract ) monoids @xmath77 .
observe that if @xmath72 is an extendible character and @xmath78 is an extension of @xmath72 , we can suppose that @xmath79 .
then @xmath80 for all @xmath81 , and thus @xmath78 is an extension of @xmath72 to a regular function of @xmath3 .
let @xmath3 be an affine algebraic monoid and @xmath43 a finite rational @xmath3-module .
for every @xmath82 and @xmath83 we define @xmath84 as @xmath85 for all @xmath86 .
a function of this form is called an _ @xmath43representative function _ or simply a _ representative function .
_ let @xmath1 be an algebraic group and let @xmath87 be a character .
let @xmath43 be a right @xmath1-module , with action @xmath88 , @xmath89 .
then the _ twisted representation _
@xmath90 is defined as follows . as a vector space @xmath91 , and the action @xmath92 is given by @xmath93 , for every @xmath34 , @xmath94 .
the following theorem is a straightforward generalization of the corresponding ( well known ) result for algebraic groups ( see ( * ? ? ?
* theorem 7.2.3 ) ) .
we include a proof for the sake of completeness .
[ theo : h1 ] let @xmath3 be an affine algebraic monoid with unit group @xmath1 , @xmath95 a closed subgroup and @xmath43 a finite dimensional rational right @xmath0-module .
there exists a finite dimensional rational right @xmath3-module @xmath96 , an extendible character @xmath97 and an injective morphism @xmath98 .
given @xmath43 as above , we proceed as in the proof of ( * ? ? ?
* theorem 7.2.3 ) .
it is well known that there exists an injective morphism of @xmath0-modules @xmath99 $ ] , where @xmath100 is a finite set of indexes ( see for example ( * ? ? ?
* theorem 4.3.13 ) ) . consider the @xmath0-morphism @xmath101 \to \bigoplus_{i } \bbbk [ h]$ ] , where @xmath102 \to \bbbk [ h]$ ] is the canonical projection . call @xmath103 $ ] and @xmath104 the restriction of @xmath105 to the @xmath0submodule @xmath106 : @xmath107_-{\beta } \ar@{^{(}->}[r ] & \bigoplus_{i } \bbbk [ m ] \ar@{->>}[d]^-{\alpha } \\ v \ar[r]_-{\theta } & \bigoplus_{i } \bbbk [ h ] } \ ] ] let @xmath108 be a finite @xmath11-linear basis of @xmath43 and let @xmath109 be a finite set such that @xmath110 .
call @xmath111 the finite dimensional @xmath1-submodule of @xmath112 $ ] generated by @xmath113 ; then @xmath111 is an @xmath3-module .
let @xmath114 be finite dimensional @xmath0-submodule of @xmath115 ( contained in @xmath106 ) generated by @xmath116 . in this way
, we produce an exact sequence of rational @xmath0modules @xmath117 . call @xmath118 and consider the commutative diagram below @xmath119 & u \otimes \bigwedge^nu \ar[r ] \ar[d ] & s \otimes \bigwedge^nu \ar[r ] \ar[d ] & v \otimes \bigwedge^nu \ar@{.>}[dl]^-{\varphi } \ar[r ] & 0 \\ 0 \ar[r ] & u \land \bigwedge^nu=0 \ar[r ] & s \land \bigwedge^nu & & } \def\objectstyle{\textstyle } \def\labelstyle{\textstyle}\ ] ] where all the solid arrows are the canonical ones , the first row is exact and in the second row the term @xmath120 equals zero by dimensional reasons notice that all the exterior products are taken inside the exterior algebra @xmath121 .
one can now prove that @xmath122 is bijective ( see the proof of ( * ? ? ?
* theorem 7.2.3 ) ) . as @xmath114 and @xmath123
are @xmath0-submodules of @xmath115 , we can view @xmath124 as an injective @xmath0morphism @xmath125 .
call @xmath72 the rational character associated to the one dimensional @xmath0-module @xmath126 , i.e. the character defined by the formula @xmath127 for all @xmath128 .
the map @xmath129 , @xmath130 , satisfies that for all @xmath128 , @xmath131 hence , if we let @xmath132 and @xmath133 , the proof of the theorem will be complete once we prove that the character @xmath72 is extendible . consider the @xmath3module @xmath134 and let @xmath135 be such that @xmath136 .
then the representative function @xmath137 $ ] is an extension of @xmath72 . indeed , if @xmath128 and @xmath138 , then @xmath139 and latexmath:[$(\alpha of @xmath72 .
let @xmath3 be an affine algebraic monoid with unit group @xmath1 .
many results about extendible characters for subgroups of @xmath1 apply to the situation of closed subgroups of @xmath3 .
[ lema:1 ] let @xmath3 be an affine algebraic monoid with unit group @xmath1 , and @xmath95 a subgroup , closed in @xmath3 .
\(1 ) let @xmath141 be an extendible character and let @xmath142 $ ] be a semiinvariant of weight @xmath56 .
consider the left action @xmath143\to \bbbk[m]$ ] given by @xmath144 , @xmath145 .
then for any @xmath86 , @xmath146 is also a semiinvariant of weight @xmath56 .
\(2 ) if @xmath147 \rightarrow \bbbk[h]$ ] is the canonical projection and @xmath148 is an extendible character , there exists an extension @xmath142 $ ] of @xmath56 such that @xmath149 .
\(3 ) a character @xmath150 is extendible if and only if there exists a rational right @xmath3module @xmath43 and an injective morphism of right @xmath0modules @xmath151 .
in other words @xmath56 is extendible if and only if there exists a rational right @xmath3module @xmath43 and a non - zero element @xmath83 such that @xmath152 for all @xmath153 .
moreover , the right @xmath3module @xmath43 can be taken to be a finite dimensional right @xmath3submodule of @xmath154 $ ] .
\(4 ) for any @xmath155 there exists @xmath156 such that @xmath157 . many of these results are straightforward generalizations of the corresponding results for algebraic groups
. see ( * ? ? ?
* lemma 7.2.8 ) . for the sake of completeness
we include the proofs of ( 3 ) and ( 4 ) . to prove ( 3 ) suppose that @xmath56 is extendible , let @xmath158 $ ] be an extension and call @xmath96 the rational right @xmath1submodule of @xmath4 $ ] generated by @xmath78 .
then the map @xmath159 , @xmath160 , does the job .
conversely , if one has an injective morphism @xmath161 and call @xmath162 , then @xmath163 .
take now @xmath164 such that @xmath165 and consider @xmath166 .
then , @xmath167 for all @xmath128 , and @xmath168 . to prove ( 4 ) consider @xmath169 . by theorem [ theo : h1 ]
there exists an extendible character @xmath56 , a right @xmath3module @xmath96 and an injective @xmath0morphism @xmath170 .
call @xmath171 , then for any @xmath128 we have that @xmath172 .
from ( 3 ) we conclude that @xmath173 .
observable subgroups were introduced by bialynicki - birula , hochschild and mostow in @xcite . since then they have been researched extensively , notably by f. grosshans .
see @xcite for a survey on this topic .
this basic notion has recently been generalized and reformulated in the context of geometric invariant theory by the authors .
see @xcite .
[ defi : obsgrp ] let @xmath1 be an affine algebraic group and let @xmath95 be a closed subgroup . the subgroup @xmath0 is _ observable _ in @xmath1 if and only if for any nonzero @xmath0-stable ideal @xmath174 $ ] we have that @xmath175 , where we consider the right action of @xmath0 on @xmath176 $ ] given by @xmath177 , for all @xmath178 $ ] @xmath179 and @xmath180 .
analogously , one can define the notion of a _
right observable subgroup _ , by considering the action @xmath181\to\bbbk[g]$ ] , @xmath182 for @xmath178 $ ] , @xmath180 , @xmath179 .
\(1 ) if @xmath183 is a closed unipotent subgroup , then @xmath123 is observable , since any @xmath123-module has non - zero invariant elements .
\(2 ) if @xmath95 is a normal closed subgroup , then @xmath0 is observable .
this follows from condition ( 2 ) of theorem [ thm : obsercond ] below .
\(3 ) let @xmath95 be a closed subgroup , such that @xmath184 .
then @xmath0 is observable in @xmath1 .
this follows for example from condition ( 6 ) of theorem [ thm : obsercond ] below .
we now present a collection of equivalent definitions of observability .
the proofs can be found in ( * ? ? ?
10.2.9 and 10.5.5 ) .
we give a different proof for the fact that @xmath0 is observable in @xmath1 if and only if @xmath185 is a group ( equivalence ( 1 ) @xmath186 ( 7 ) of theorem [ thm : obsercond ] ) .
this will provide some insight into the more general situation of algebraic monoids given in theorem [ theo:1 ] .
[ thm : obsercond ] let @xmath1 be an affine algebraic group and @xmath95 a closed subgroup . then the following conditions are equivalent : 1 .
the subgroup @xmath0 is observable in @xmath1 .
2 . the homogeneous space @xmath2 is a quasi affine variety .
3 . for an arbitrary proper and closed subset @xmath187
, there exists an non - zero invariant regular function @xmath188^h$ ] such that @xmath189
. moreover , if @xmath1 is connected the above conditions are equivalent to any of the following .
1 . @xmath190^h\bigr\}$ ] .
@xmath191\bigr]^h= \bigl[\bbbk [ g]^h\bigr]$ ] .
3 . @xmath192 .
4 . @xmath185 is a group . to prove that ( 1 ) @xmath193 ( 7 ) , let @xmath0 be observable in @xmath1 and @xmath194 an extendible character
let @xmath178 $ ] be a semiinvariant of weight @xmath72 .
then the ideal @xmath195 $ ] is @xmath0-stable , and hence there exists @xmath196 $ ] such that @xmath197 for all @xmath75 .
it follows that @xmath198 for all @xmath75 .
thus @xmath199 . to prove that ( 7 ) @xmath193 ( 1 ) , let @xmath200 $ ] be a non - zero @xmath0-stable ideal , and consider @xmath201 $ ] .
then @xmath202 is normal in @xmath0 , with @xmath203 , and such that @xmath204 is a torus ; in particular @xmath205 , since @xmath203 .
it follows that @xmath202 is observable in @xmath1 , and thus @xmath206 is a @xmath0-module .
then @xmath207 is a right @xmath208-module , and hence @xmath209 hence , there exists @xmath210 and @xmath211 such that @xmath212 for all @xmath75 . thus @xmath213 . since @xmath185 is a group , then @xmath199 and hence there exists @xmath214 $ ] such that @xmath215 for all @xmath75 .
it follows that @xmath216 .
[ obser.def ] let @xmath1 be an affine algebraic group acting on an affine variety @xmath8 .
we say that the action is _ observable _ if for any nonzero @xmath1-stable ideal @xmath217 $ ] we have that @xmath218 .
\(1 ) the action of an unipotent group on an affine variety is always observable , since for every right @xmath123-module @xmath3 , we have that @xmath219 .
\(2 ) let @xmath1 be an algebraic group and @xmath0 a closed subgroup .
then the action of @xmath0 on @xmath1 by left translations is observable if and only if @xmath0 is observable in @xmath1 , in the sense of definition [ defi : obsgrp ] , if and only if the action by right translations is observable .
observable actions have been studied in detail in @xcite , where several characterizations of this important property were presented .
we collect here the ones that we need in what follows .
we begin by recalling some notation .
let @xmath1 be an affine group acting on an affine variety @xmath8 .
we let @xmath220 that is , @xmath221 is the set of orbits of maximal dimension that are closed .
the reader should be aware that @xmath221 can be empty .
[ thm : obserchar ] let @xmath1 be a connected affine algebraic group acting on an irreducible affine variety @xmath8
. then the following are equivalent 1 .
the action is observable .
2 . 1 . @xmath222^g\bigr]=\bigl[\bbbk[x]^g\bigr]$ ] and 2 .
@xmath221 contains a non - empty open subset .
there exists a nonzero invariant @xmath223^g$ ] such that the action of @xmath1 on @xmath224 is observable
. see ( * ? ? ?
* proposition 3.2 and theorem 3.10 ) .
if @xmath1 is a reductive group acting on an affine variety @xmath8 , then a result of popov ( see ( * ? ? ?
* theorem 4 ) ) guarantees that if @xmath221 is nonempty then it is an open set . in @xcite
this result was used to prove that in this case , condition ( 2ii ) of theorem [ thm : obserchar ] is sufficient to guarantee observability .
[ prop : obsred ] let @xmath1 be a reductive group acting on an affine variety @xmath8 . then the action is observable if and only if @xmath225
. see ( * ? ? ?
* theorem 4.7 ) .
let @xmath1 be an affine algebraic group acting on the affine variety @xmath8 .
it is well known that the categorical quotient does not necessarily exist , even when @xmath36^g$ ] is finitely generated . however
, if @xmath36^g$ ] is finitely generated , then @xmath226^g\bigr)$ ] satisfies a universal property in the category on the affine algebraic varieties .
let @xmath1 be an affine algebraic group acting on an affine variety @xmath8 , in such a way that @xmath36^g$ ] is finitely generated .
the _ affinized quotient _ of the action is the morphism @xmath227^g\bigr)$ ] .
it is clear that @xmath228 satisfies the following universal property .
_ let @xmath229 be an affine variety and @xmath230 a morphism constant on the @xmath1-orbits .
then there exists a unique @xmath231 such that the following diagram is commutative . _ indeed , it is clear that the induced morphism @xmath232\to \bbbk[x]$ ] factors trough @xmath36^g$ ] .
[ rem : quotnotsurj ] it is clear that the morphism @xmath233 is dominant .
however , @xmath228 is not necessarily surjective .
for example , consider a semisimple group @xmath234 and its maximal unipotent subgroup @xmath1 acting on @xmath8 by left translation .
we now adapt the notion of observability to the situation of subgroups of algebraic monoids .
[ defi : obsmon ] let @xmath3 be an algebraic monoid with unit group @xmath1 , and let @xmath95 be a closed subgroup .
we say that @xmath0 is _
( left ) observable in @xmath3 _ if the action by left multiplication @xmath235 , @xmath236 , is observable in the sense of definition [ obser.def ] .
similarly , we say that @xmath0 is _ right observable in @xmath3 _ if the action by right multiplication @xmath237 , @xmath238 , is observable .
[ exam : obsmon ] ( 1 ) since the action of a unipotent group on an affine variety is observable ( all orbits are closed ) , it follows that any unipotent subgroup @xmath123 of @xmath49 is observable in @xmath3 .
\(2 ) if @xmath239 is an algebraic group , then definitions [ defi : obsgrp ] and [ defi : obsmon ] coincide .
\(3 ) if follows from proposition [ prop : obsred ] that if @xmath70 is reductive , and closed in @xmath3 , then @xmath0 is observable in @xmath3 . indeed , if @xmath240 , then @xmath241 .
[ thm : istabcond ] let @xmath3 be an affine irreducible algebraic monoid with unit group @xmath1 , and let @xmath242 be an observable subgroup .
then \(1 ) the subgroup @xmath0 is observable in @xmath1 .
\(2 ) @xmath0 closed in @xmath3 .
\(3 ) @xmath243\bigr]^h=\bigl [ \bbbk [ g]\bigr]^h= \bigl[\bbbk [ g]^h\bigr]= \bigl[\bbbk [ m]^h\bigr]$ ] .
\(4 ) the subgroup @xmath0 satisfies @xmath244^h\bigr\}= \\ & \ \bigl\{x \in g\mathrel { : } f\cdot x = f , \ \forall f \in \bbbk [ m]^h\bigr\}=\\ & \
\bigl\{x \in m\mathrel { : } f\cdot x = f , \ \forall f \in \bbbk [ m]^h\bigr\}^0 . \end{split}\ ] ] recall that if @xmath245 is an algebraic monoid , then @xmath246 is the unique irreducible component containing @xmath45 , see for example ( * ? ? ?
* thm . 4 ) .
\(1 ) let @xmath247 be a @xmath0-stable closed subset .
then @xmath248 is @xmath0-stable , and it follows that there exists a @xmath249 $ ] . since @xmath1 is open in @xmath3 , it follows that @xmath250 $ ] .
\(2 ) since @xmath0 is observable in @xmath3 , it follows from theorem [ thm : obserchar ] that @xmath251 contains a non - empty open subset . since @xmath252 , @xmath1
is contained in @xmath253 , and hence there exists @xmath34 such that @xmath254 , i.e. such that @xmath255 is closed in @xmath3 . since multiplication by an element of @xmath256 is an isomorphism , it follows that @xmath0 is closed in @xmath3 .
\(3 ) first observe that @xmath257^h\bigr]\subset \bigl [ \bbbk [ m]\bigr]^h=\bigl [ \bbbk [ g]\bigr]^h$ ] .
let @xmath258\bigr]^h$ ] , and consider the ideal @xmath259\mathrel { : } fg\in\bbbk[m]\bigr\}$ ] .
then @xmath100 is a non - zero @xmath0-stable ideal , and hence there exists @xmath260 .
it follows that @xmath261^h$ ] and hence @xmath262^h\bigr]$ ] .
the remaining equality follows from theorem [ thm : obsercond ] .
\(4 ) the first equality follows from theorem [ thm : obsercond ] .
it is clear that @xmath263^h\bigr\ } \subset b=\bigl\{x \in g\mathrel { : } f\cdot x = f , \
\forall f \in \bbbk [ m]^h\bigr\}$ ] .
let @xmath264 and @xmath178^h$ ] . by ( 3 )
, it follows that @xmath265^h\bigr]=\bigl[\bbbk[m]^h\bigr]$ ] .
hence , there exist @xmath266^h$ ] such that @xmath267 .
it follows that @xmath268 . in order to prove the last equality
, we first observe that @xmath269^h\bigr\ } \bigr)$ ] is a closed submonoid of @xmath3 .
since @xmath1 is open in @xmath3 , it follows that @xmath270 [ theo:1 ] let @xmath3 be an irreducible affine algebraic monoid with unit group @xmath1 and let @xmath95 be a subgroup , closed in @xmath3 .
then the following conditions are equivalent .
\(1 ) the subgroup @xmath0 is observable in @xmath3 .
\(2 ) @xmath76 is a group .
that is , for every @xmath271 , @xmath272 .
\(3 ) @xmath273 , i.e. every rational character is extendible .
\(4 ) for every finite dimensional rational right @xmath0-module @xmath43 there exists a finite dimensional rational right @xmath3-module @xmath96 and an injective morphism of @xmath0-modules @xmath274 .
\(5 ) @xmath0 is observable in @xmath1 and for any ( some ) determinant @xmath275 , then @xmath276 .
in order to prove that ( 1 ) implies ( 2 ) , assume that @xmath0 is observable in @xmath3 , and let @xmath277 be an extendible character .
let @xmath256 be an extension of @xmath72 .
then the ideal @xmath278 $ ] is a nonzero @xmath0-stable ideal and hence there exists @xmath66 $ ] such that @xmath279 for all @xmath75 .
it follows that @xmath280 for all @xmath75 .
that is , @xmath281 .
since @xmath76 generates @xmath282 as a group , it is clear that conditions ( 2 ) and ( 3 ) are equivalent . to prove that ( 2 ) implies ( 1 ) , assume that @xmath76 is a group and let @xmath283 $ ] be a non - zero @xmath0-stable ideal .
let @xmath201 $ ] , as in the proof of theorem [ thm : obsercond ] .
then @xmath202 is observable in @xmath0 .
consider a determinant @xmath284 .
since @xmath285 , it follows that @xmath286 .
thus , we have found an @xmath202-invariant regular function @xmath287^{h_{\operatorname{uni } } } $ ] , such that @xmath176=\bbbk[m]_{\det}$ ] .
it follows from theorem [ thm : obserchar ] that @xmath202 is observable in @xmath3 , and thus the right @xmath0-module @xmath207 is not trivial .
then , as in the proof of theorem [ thm : obsercond ] , it follows that @xmath207 is a right @xmath288-module , and hence @xmath289 hence , there exists @xmath210 and @xmath211 such that @xmath212 for all @xmath75 .
thus @xmath277 .
since @xmath76 is a group , it follows that @xmath281 .
thus , there exists @xmath214 $ ] such that @xmath215 for all @xmath75 .
it follows that @xmath216 . to prove that ( 2 ) implies ( 4 ) , we follow the idea of the proof of ( *
* theorem 10.2.9 ) .
if @xmath43 is a finite dimensional rational right @xmath0-module , using theorem [ theo : h1 ] we deduce the existence of an extendible character @xmath56 , a finite dimensional rational right @xmath3-module @xmath96 and an injective map @xmath290 such that @xmath291 for all @xmath128 . by hypothesis the character
@xmath292 is extendible , so if we take @xmath158 $ ] that extends @xmath292 and call @xmath293 the right @xmath3-submodule of @xmath294 $ ] generated by @xmath78 , we can define an injective map @xmath295 , given as @xmath296 . if we endow @xmath297 with the diagonal right @xmath3-module structure
, the following computation shows that @xmath298 is @xmath0-equivariant .
let @xmath128 .
then @xmath299 we now prove that ( 4 ) implies ( 3 ) .
let @xmath300 be a rational character of @xmath0 and consider the rational @xmath0-module @xmath301 .
if @xmath298 and @xmath96 are as in condition ( 4 ) we conclude , using lemma [ lema:1 ] , that @xmath302 is extendible .
assume now that ( 1 ) holds .
then by theorem [ thm : istabcond ] , @xmath0 is observable in @xmath1 , and it follows from ( 2 ) that @xmath303 .
hence , ( 5 ) holds .
finally , if ( 5 ) holds , then by theorem [ thm : obsercond ] , @xmath304 .
let @xmath305 be a determinant such that @xmath276 , and let @xmath66 $ ] be an extension of @xmath306 .
if @xmath71 , let @xmath196 $ ] be an extension of @xmath72 . then there exists @xmath307 $ ] and @xmath54 such that @xmath308 .
therefore , for every @xmath75 and @xmath180 , we have that @xmath309 it follows that @xmath310 ; that is @xmath307 $ ] is an extension of @xmath311
. then @xmath312 $ ] is an extension of @xmath72 .
[ thm : obserrep ] let @xmath3 be an irreducible affine algebraic monoid with unit group @xmath1 and let @xmath95 be a subgroup , closed in @xmath3
. then the following conditions are equivalent .
\(1 ) the subgroup @xmath0 is observable in @xmath3 .
\(2 ) there exists a finite dimensional rational right @xmath3-module @xmath43 and @xmath9 such that @xmath313 . in particular ,
if condition ( 2 ) holds , then @xmath314 is closed in @xmath3 .
to prove that condition ( 1 ) implies condition ( 2 ) , we adapt the proof of the case @xmath315 ( see ( * ? ? ?
* corollary 7.3.6 ) ) .
first we observe that , since @xmath1 is observable in @xmath3 , @xmath0 is observable in @xmath1 and @xmath316^h\bigr]= \bigl[\bbbk [ g]\bigr]^h= \bigl[\bbbk [ m]\bigr]^h=\bigl[\bbbk [ m]^h\bigr]$ ] .
let @xmath317^h$ ] be such that @xmath318 generates @xmath319\bigr]^h$ ] over @xmath42 .
let @xmath320 $ ] be the finite dimensional rational right @xmath3-submodule generated by @xmath321 .
let @xmath322 , @xmath323 and consider the stabilizer @xmath324 .
it is clear that @xmath325 .
conversely , if @xmath326 , then @xmath327 , @xmath328 , and hence @xmath329 , @xmath330 . as the elements @xmath331 , @xmath330 , generate @xmath257\bigr]^h$ ] , we conclude that @xmath332 for all @xmath333\bigr]^h=\bigl[\bbbk [ g]\bigr]^h$ ] .
hence , @xmath334 , as follows by example from ( * ? ? ?
* corollary 7.3.4 ) . to prove that ( 2 ) implies ( 1 ) , let @xmath335 be as in ( 2 ) and consider the following commutative diagram since @xmath2 is quasi - affine , it follows that @xmath0 is observable in @xmath1 , and hence @xmath336^h\bigr]=\bigl [ \bbbk[g]\bigr]^h$ ] .
thus , @xmath337\bigr]=\bbbk(v\cdot
g)\cong \bigl[\bbbk[g]^h\bigr]= \bigl [ \bbbk[g]\bigr]^h.\ ] ] on the other hand , the orbit morphism @xmath338 is dominant , and hence induces an inclusion @xmath339\hookrightarrow \bbbk[m]$ ] . since @xmath313
, it follows that @xmath340 and hence @xmath341\bigr ) \subset \bbbk[m]^h$ ] .
thus , @xmath342\bigr]^h = \bigl [ \bbbk[g]\bigr]^h= \bigl[\bbbk[\overline{v\cdot m}]\bigr]\subset \bigl[\bbbk[m]^h\bigr]\subset \bigl [ \bbbk[m]\bigr]^h,\ ] ] and hence the action by right multiplication of @xmath0 on @xmath3 is observable . finally , recall that observable subgroups are closed in @xmath3 .
[ rem : q1 ] it is clear that all the results of this section remain valid when considering right observability simply by adapting the statements and proofs in an obvious way . if @xmath239 is an algebraic group , it is well known that @xmath95 is ( left ) observable in @xmath1 if and only if @xmath0 is right observable .
indeed , in this case the antipode @xmath343\to \bbbk[g]$ ] , @xmath344 , induces an isomorphism @xmath176^h\cong { } ^h\bbbk[g]$ ] . in the more general case of an algebraic monoid @xmath3
, this line of reasoning can not be applied , since @xmath345\bigl)$ ] is not included in @xmath154 $ ] .
this raises the following question . *
* _ let @xmath3 be an algebraic monoid with unit group @xmath1 , and @xmath95 a closed subgroup , ( left ) observable in @xmath3 .
is @xmath0 right observable in @xmath3 ? _
let @xmath3 be an algebraic monoid of unit group @xmath1 and let @xmath95 a closed subgroup , observable in @xmath3 .
our goal is to prove that if @xmath0 is normal in @xmath1 , then the affinized quotient @xmath7 is an affine monoid with unit group @xmath2 .
see theorem [ thm : normalimplisobse ] . before doing so , we present a general result about the affinized quotient of a monoid by an observable subgroup .
let @xmath3 be an algebraic monoid with unit group @xmath1 and let @xmath95 be a closed subgroup , observable in @xmath3 .
assume that @xmath154^h$ ] is finitely generated .
then @xmath7 is an affine embedding of @xmath2 .
that is , @xmath7 is an affine @xmath1-variety , with an open orbit isomorphic to @xmath2 . where the existence of @xmath124 follows from the universal property of the quotient .
since @xmath0 is observable in @xmath3 , it follows from theorem [ thm : istabcond ] that @xmath346^h\bigr]= \bigl [ \bbbk [ m]\bigr]^h=\bigl [ \bbbk [ g]\bigr]^h= \bigl[\bbbk [ g]^h\bigr]$ ] .
hence , @xmath124 is a birational dominant @xmath1-morphism .
since @xmath2 is a homogeneous @xmath1-space , it follows that @xmath124 is an open immersion .
[ prop : quotbynormred ] let @xmath3 be a reductive monoid with unit group @xmath1 , and let @xmath95 be a normal subgroup , closed in @xmath3
. then the categorical quotient @xmath347 is an affine algebraic monoid with unit group @xmath2 .
furthermore , @xmath348 is a morphism of algebraic monoids .
first of all , we prove that @xmath0 is observable in @xmath3 .
indeed , @xmath0 is a normal subgroup of @xmath1 and thus it is a reductive group .
since @xmath242 is closed , it follows that @xmath349 is closed in @xmath3 for all @xmath34 , and we deduce from proposition [ prop : obsred ] that @xmath0 is observable in @xmath3 .
let @xmath350\to \bbbk[m]$ ] , @xmath351 @xmath29 , @xmath138 , be the canonical action . if @xmath66^h$ ] and @xmath352 , then @xmath353 where @xmath354 is such that @xmath355 .
if follows that @xmath154^h$ ] is a @xmath356-submodule of @xmath154 $ ] .
let now @xmath357
. then @xmath358 .
indeed , if @xmath34 , then @xmath359 for some @xmath354 , and hence @xmath360 . in other words , @xmath361 , and thus @xmath362
it follows that @xmath363 .
hence , @xmath347 is an affine @xmath364-variety and the coproduct @xmath365\to \bbbk[m]\otimes \bbbk[m]$ ] is such that @xmath366^h\bigr ) \subset \bigl(\bbbk[m]^h\otimes \bbbk[g]\bigr)\cap \bigl(\bbbk[g ] \otimes\bbbk[m]^h\bigr)=\bbbk[m]^h\otimes \bbbk[m]^h.\ ] ] moreover , since @xmath154^h\subset \bbbk[g]^h=\bbbk[g / h]$ ] , it follows that @xmath347 is an algebraic monoid and that we have a commutative diagram where the existence of @xmath124 follows from the universal property of the quotient .
then @xmath124 is a dominant morphism of algebraic monoids , and hence @xmath367 .
since @xmath0 is observable in @xmath3 it follows that @xmath368\bigr]^h=\bigl[\bbbk[m]\bigr]^h= \bigl[\bbbk[m]^h\bigr ] = \bbbk(m//h).\ ] ] thus , @xmath369 is an injective birational @xmath1-morphism , and it follows that @xmath370 is an open immersion .
hence , @xmath124 is an isomorphism .
[ thm : passtored ] let @xmath3 be an affine algebraic monoid with unit group @xmath1 . then @xmath154^{r_u(g)}$ ] is a finitely generated algebra .
moreover , @xmath371^{r_u(g)}\bigr)$ ] is an affine algebraic monoid with unit group @xmath372 .
the morphism @xmath373 is a surjective morphism of algebraic monoids , and satisfies the following universal property . for any morphism @xmath374 , of algebraic monoids , such that @xmath375
there exists a unique morphism @xmath376 such that the diagram it follows from ( * ? ? ? * theorem 3.18 ) that @xmath240 if and only if @xmath379 .
but by ( * ? ? ?
* corollary 6.10 ) , @xmath380 , where @xmath381 is a maximal tours of @xmath0 .
hence , @xmath382 if and only if @xmath383 for any ( some ) maximal torus @xmath381 . by ( * ? ? ?
* theorem 2.5 ) , it follows that , if @xmath384 is a maximal torus of @xmath1 , then @xmath385 is an isomorphism .
let now @xmath381 be a maximal torus of @xmath0 and @xmath384 a maximal torus of @xmath1 such that @xmath386 .
then @xmath387 is an isomorphism .
in particular , @xmath383 if and only if @xmath388 .
since @xmath389 is a maximal tours , it follows that @xmath382 if and only if @xmath390 .
we conclude the proof by showing that the affinized quotient @xmath391 is a surjective morphism of algebraic monoids ( this fact is implicit in @xcite but was not stated or proved ) . by construction
, @xmath228 is a morphism of algebraic monoids .
since @xmath392 is a reductive monoid , it follows that @xmath393 see for example ( * ? ? ?
* theorem 4.2 ) .
the surjectivity of @xmath228 follows now from the fact that @xmath394 .
[ thm : normalimplisobse ] let @xmath3 be an algebraic monoid with unit group @xmath1 , and let @xmath95 be a normal subgroup , closed in @xmath3
. then the action @xmath395 is observable .
moreover , if @xmath154^h$ ] is finitely generated , then the affinized quotient @xmath7 is an affine algebraic monoid , with unit group @xmath2 .
let @xmath373 be the affinized quotient as in theorem [ thm : passtored ] . by [ ibid ]
@xmath392 is a reductive algebraic monoid with unit group @xmath372 .
we use @xmath228 to help us find a determinant function on @xmath3 ( the function @xmath396 below ) with suitable properties . consider the normal subgroup @xmath397 . then @xmath398 is closed in @xmath392 and hence by proposition [ prop : quotbynormred ] it follows that @xmath399 is an affine algebraic monoid .
let @xmath400 be the affinized quotient and @xmath401 a character such that @xmath402 .
then @xmath403 is a character such that @xmath404 and @xmath405 .
in particular @xmath406^h$ ] .
since @xmath0 is normal in @xmath1 @xmath0 is observable in @xmath1 . in other words
, the action of @xmath0 on @xmath407 is observable .
it follows from theorem [ thm : obserchar ] that the action of @xmath0 on @xmath3 is observable . to prove the last assertion we use the same arguments as in the proof of proposition [ prop : quotbynormred ] .
since @xmath408^h\bigr ) \subset \bbbk[m]^h\otimes \bbbk[m]^h$ ] , it follows that @xmath7 is an algebraic monoid and that we have a commutative diagram where the existence of @xmath124 follows from the universal property of the quotient .
then @xmath124 is a dominant morphism of algebraic monoids , and hence @xmath409 .
since the action of @xmath0 on @xmath3 is observable , it follows from theorem [ thm : istabcond ] that @xmath410^h\bigr]=\bigl[\bbbk[m]\bigr]^h$ ] , and hence @xmath124 is a birational morphism . since @xmath56 is @xmath1-equivariant , to follows that @xmath124 is an open immersion , and thus @xmath411 .
let @xmath245 be an algebraic monoid and @xmath374 a morphism of algebraic monoids such that @xmath412 .
then there exists a morphism of algebraic monoids @xmath413 such that the following diagram commutes .
be the chevalley decomposition of @xmath245 , where @xmath105 is the albanese morphism ( see ( * ? ? ?
* theorem 3.2.1 ) and ( * ? ? ? * theorem 1.1 ) ) . since @xmath3 is affine , if follows that form the property of the albanese morphism that @xmath414 ( see for example ( * ? ? ?
* theorem 5.1 ) ) .
since @xmath78 is a morphism of algebraic monoids , such that @xmath415 , it follows that @xmath78 is constant on the @xmath0-orbits , and hence @xmath416\to \bbbk[m]$ ] factors through @xmath154^h$ ] . in other words
, there exists @xmath417 such that @xmath418 .
[ exam : normalimp ] let @xmath419 then @xmath0 is observable in @xmath420 and closed in @xmath3 , but there exists no character @xmath421 such that @xmath422 and @xmath423 .
since @xmath0 is closed in @xmath1 and since @xmath424 , @xmath425 , is an isomorphism , it follows that @xmath426 is closed in @xmath3 for all @xmath427 .
since @xmath0 is reductive , it follows from proposition [ prop : obsred ] that the action of @xmath0 on @xmath3 is observable .
[ rem : openquest ] example [ exam : normalimp ] shows that the condition of @xmath95 being a normal subgroup is crucial in theorem [ thm : normalimplisobse ] .
however , in that example the action of @xmath0 on @xmath3 is observable .
this raises the following question .
if q2 has a positive answer , then q1 ( see remark [ rem : q1 ] ) has a positive answer .
indeed , assume that q2 has a positive answer and let @xmath242 be a left observable subgroup .
then by theorem [ thm : istabcond ] it follows that @xmath0 is observable in @xmath1 , and hence @xmath0 is right observable in @xmath1 .
if then follows from q2 that @xmath0 is right observable in @xmath3 . | a closed subgroup @xmath0 of the affine , algebraic group @xmath1 is called _ observable _ if @xmath2 is a quasi - affine algebraic variety . in this paper
we define the notion of an observable subgroup of the affine , algebraic _ monoid _ @xmath3 .
we prove that a subgroup @xmath0 of @xmath1 is observable in @xmath3 if and only if @xmath0 is closed in @xmath3 and there are `` enough '' @xmath0-semiinvariant functions in @xmath4 $ ] .
we show also that a closed , normal subgroup @xmath0 of @xmath1 ( the unit group of @xmath3 ) is observable in @xmath3 if and only if it is closed in @xmath3 . in such a case there
exists a _ determinant _
@xmath5 such that @xmath6 . as an application
, we show that in this case the _
affinized quotient _ @xmath7 of @xmath3 by @xmath0 is an affine algebraic monoid scheme with unit group @xmath2 . |
the present analysis was made in a framework of svd-2 experiment , the main goal of which is a study of the charm hadroproduction at the near - threshold energy@xcite .
svd-2 setup consists of high - precision microstrip vertex detector ( beam telescope , active target and tracking detector ) , large aperture magnetic spectrometer , multicell threshold cherenkov counter and cerenkov full absorption lead glass gamma detector .
the primary vertex position determination procedure was based on well - known `` tear - down '' approach @xcite . only events with a well reconstructed primary vertex were selected . after excluding the tracks that belongs to primary vertex , the secondary vertex position was determined by finding v0-decay downstream the primary vertex .
the primary vertex resolution was estimated as @xmath6 for z - coordinate and @xmath7 for x(y)-coordinates .
for the two - tracks secondary vertices ( @xmath8 ) those values were @xmath9 and @xmath10 respectively .
the impact parameter resolution for @xmath11 momentum tracks is about @xmath12 .
the angular acceptance of the vertex detector averages to @xmath13 .
the svd-2 setup permits obtaining the high effective mass resolution of @xmath14 for @xmath15 and @xmath16 for @xmath17 masses ( see fig.[k0_lambda_mass ] ) .
the momentum resolution for the track with 15 measured hits is @xmath18 in the @xmath19 momentum range .
the angular measurement error was estimated to be @xmath20 .
the angular acceptance of spectrometer averages to @xmath21 for horizontal and @xmath22 for vertical coordinates .
invariant mass spectrum .
a window corresponding to @xmath23 is shown by the vertical lines .
right : the @xmath24 invariant mass spectrum.,scaledwidth=80.0% ] invariant mass spectrum . a window corresponding to @xmath23 is shown by the vertical lines .
right : the @xmath24 invariant mass spectrum.,scaledwidth=80.0% ] the combined @xmath25 and @xmath26 invariant mass spectrum is shown on fig.[sigma_ks_mass]a .
the @xmath27 peak is clearly seen on the distribution .
fig.[sigma_ks_mass]b shows @xmath28 invariant mass spectrum .
@xmath29 peak is clearly seen .
the masses of observed @xmath15 , @xmath17 and also masses and widths of @xmath27 and @xmath29 are consistent with their pdg values@xcite . and @xmath26 invariant mass spectrum .
b ) @xmath28 invariant mass spectrum.,scaledwidth=40.0% ]
exotic baryons with 5-quarks content ( pentaquarks ) and their properties have been predicted by diakonov , petrov , and polyakov @xcite in the framework of the chiral soliton model , although such 5-quarks structures were proposed years ago @xcite .
the lightest member of the pentaquarks antidecuplet , @xmath0-baryon , has positive strangeness , mass @xmath30 , @xmath31 , spin @xmath32 and even parity . experimental evidence for @xmath0-baryon with positive strangeness came recently from several experimental groups ( leps@xcite , diana - itep @xcite , clas @xcite , saphir @xcite ) . in those experiments @xmath0-baryon
was observed in the @xmath33 or @xmath1 invariant mass spectra with the mass near @xmath34 .
more recently hermes collaboration observed narrow baryon state at the mass of @xmath35 in quasi - real photoproduction@xcite .
also zeus collaboration @xcite reported an evidence of the exotic baryon in @xmath1-channel with the mass of @xmath36 .
we were searching for the @xmath0-baryon in the reaction @xmath37 , @xmath38 , @xmath39 . the data pre - selected at the search for the charm production were used .
it consisted of events with v0-candidates decayed within the vertex detector : decay length @xmath40 with a mean of @xmath41 .
the following criteria were then applied : * primary charged particles multiplicity @xmath42 . minimizes the combinatorial background , reduces the probability of appearance of the events with rescattering on nuclei and background of @xmath15-mesons produced in the central rapidity region .
* a presence of proton as well - measured primary positive track with a momentum of @xmath43 leaving no hit in cherenkov counter . * @xmath44 to select well - identified @xmath15 . *
@xmath45 , where @xmath46 is angle of flight of @xmath1-system in the center mass system of the beam proton and the target nucleon(beam proton fragmentation region ) .
* @xmath47 kinematical cut @xcite : effectively destroys most of the decays of @xmath48-resonances with high masses while cutting only @xmath49 of @xmath0-peak events . resulting distribution
is shown on fig .
[ theta ] .
the distribution was fitted by gaussian function and fourth - order polynomial background .
there are 50 events in the peak over 78 background events .
the statistical significance for the fit inside a @xmath50 mass window was calculated as @xmath51 , where @xmath52 is the number of counts in the background fit under the peak and @xmath53 is the number of counts in the peak .
we estimate the significance to be of @xmath5 .
it is impossible to determine the strangeness of this state in such an inclusive reaction , however there are no reported @xmath48-resonances in @xmath54 mass area , so we interpret observed state as recently reported @xmath0-baryon with a positive strangeness .
invariant mass spectrum in the reaction @xmath55 .
dashed line : background obtained from fritiof simulations.,scaledwidth=50.0% ] the a - dependence analysis in the observed peak area showed no difference ( within measuring errors ) from the @xmath56 dependence for background inelastic events ( fig .
[ adepend ] ) . in a summary ,
the inclusive reaction @xmath57 was studied at ihep accelerator with proton energies at @xmath58 using svd-2 detector . with several cuts
applied a narrow baryon resonance was observed with mass @xmath59 and @xmath60 .
the width of this state is close to svd-2 experimental resolution for @xmath1-system and its mass and width are consistent with recently reported @xmath0-resonance@xcite , which was predicted as an exotic pentaquark ( @xmath61 ) baryon state .
events to the total of @xmath15 events for the different target materials , scaledwidth=35.0% ]
we thank hsqcd04 organizing committee and personally v. kim for providing the excellent , warm and stimulating atmosphere during the conference and for the financial support .
we are grateful to v. kubarovsky , b. levchenko and n. zotov for useful comments and suggestions . | we report on the svd-2 experiment data analysis aimed to search for an exotic baryon state , the @xmath0-baryon , in a @xmath1 decay mode with ihep u-70 accelerator proton beam at @xmath2 .
a resonant structure with @xmath3 and @xmath4 was found in the @xmath1 invariant mass spectrum , with the statistical significance of this peak estimated as @xmath5 . |
we consider euler equations for the evolution of a fluid flowing in a pipe with varying section @xmath1 , see ( * ? ? ? * section 8.1 ) or @xcite : @xmath2 = p\left ( \rho , e\right ) \ , \partial_{x}a \\ \displaystyle \partial_{t } ( ae ) + \partial_{x } \left [ a f\left ( \rho , q , e\right ) \right ] = 0 \end{array } \right.\ ] ] where , as usual , @xmath3 is the fluid density , @xmath4 is the linear momentum density and @xmath5 is the total energy density . moreover @xmath6 with @xmath7 being the internal energy density , @xmath8 the flow of the linear momentum density and @xmath9 the flow of the energy density .
the above equations express the conservation laws for the mass , momentum , and total energy of the fluid through the pipe .
below , we will often refer to the standard case of the ideal gas , characterized by the relations @xmath10 for a suitable @xmath11 .
note however , that this particular equation of state is necessary only in case * ( p ) * of proposition [ 1prop : otherpsi ] and has been used in the examples in section [ 1sec : blow - up ] . in the rest of this work ,
the usual hypothesis ( * ? ? ?
* formula ( 18.8 ) ) , that is @xmath12 , @xmath13 and @xmath14 , are sufficient .
the case of a sharp discontinuous change in the pipe s section due to a junction sited at , say , @xmath15 , corresponds to @xmath16 for @xmath17 and @xmath18 for @xmath19 .
then , the motion of the fluid can be described by @xmath20 for @xmath21 , together with a _ coupling condition _ at the junction of the form : @xmath22 above , we require the existence of the traces at @xmath15 of @xmath23 .
various choices of the function @xmath24 are present in the literature , see for instance @xcite in the case of the @xmath25-system and @xcite for the full @xmath26 system ( [ 1eq : e ] ) . here , we consider the case of a general coupling condition which comprises all the cases found in the literature . within this setting , we prove the well posedness of the cauchy problem for ( [ 1eq : e])([1eq : junction ] ) .
once this result is obtained , the extension to pipes with several junctions and to pipes with a @xmath27 section is achieved by the standard methods considered in the literature . for the analytical techniques to cope with networks having more complex geometry
, we refer to @xcite .
the above statements are global in time and local in the space of the thermodynamic variables @xmath28 .
indeed , for any fixed ( subsonic ) state @xmath29 , there exists a bound on the total variation @xmath30 of the pipe s section , such that all sections below this bound give rise to cauchy problems for ( [ 1eq : e])([1eq : junction ] ) that are well posed in @xmath31 .
we show the necessity of this bound in the conditions found in the current literature .
indeed , we provide explicit examples showing that a wave can be arbitrarily amplified through consecutive interactions with the pipe walls , see figure [ 1fig : f ] .
the paper is organized as follows .
the next section is divided into three parts , the former one deals with a single junction and two pipes , then we consider @xmath32 junctions and @xmath33 pipes , the latter part presents the case of a @xmath27 section .
section [ 1sec : coupl cond ] is devoted to different specific choices of coupling conditions ( [ 1eq : junction ] ) . in section [ 1sec : blow - up ]
, an explicit example shows the necessity of the bound on the total variation of the pipe s section .
all proofs are gathered in section [ 1sec : tech ] .
throughout , we let @xmath34 . we denote by @xmath35 the real halfline @xmath36 , while @xmath37 0 , + \infty \right [ $ ] . following various results in the literature , such as @xcite ,
we limit the analysis in this paper to the _ subsonic _ region given by @xmath38 and @xmath39 , where @xmath40 is the @xmath41th eigenvalue of ( [ 1eq : e ] ) , see ( [ 1eigenv ] ) . without any loss of generality
, we further restrict to @xmath42 note that we fix _ a priori _ the sign of the fluid speed @xmath43 , since @xmath44 .
we now give the definition of weak @xmath45solution to the cauchy problem for ( [ 1eq : e ] ) equipped with the condition ( [ 1eq : junction ] ) , extending ( * ? ? ?
* definition 2.1 ) and ( * ? ? ?
* definition 2.2 ) to the @xmath26 case ( [ 1eq : e ] ) and comprising the particular case covered in ( * ? ? ?
* definition 2.4 ) .
[ 1def : weaksolution ] let @xmath46 , @xmath47 and two positive sections @xmath48 , @xmath49 be given .
a @xmath24-solution to ( [ 1eq : e ] ) with initial datum @xmath50 is a map @xmath51 such that 1 .
: : for @xmath21 , @xmath52 is a weak entropy solution to ( [ 1eq : e ] ) ; 2 . : : for a.e .
@xmath53 , @xmath54 ; 3 .
: : for a.e .
@xmath55 , the coupling condition ( [ 1eq : junction ] ) at the junction is met .
below , extending the @xmath56 case of the @xmath25-system , see @xcite , we consider some properties of the coupling condition ( [ 1eq : junction ] ) , which we rewrite here as @xmath57 ( @xmath580 ) : : regularity : @xmath59 .
( @xmath581 ) : : no - junction case : for all @xmath60 and all @xmath61 , then @xmath62 ( @xmath582 ) : : consistency : for all positive @xmath63 and all @xmath64 , @xmath65 moreover , by an immediate extension of ( * ? ? ?
* lemma 2.1 ) , ( @xmath66 ) ensures that ( [ 1eq : psi ] ) implicitly defines a map @xmath67 in a neighborhood of any pair of subsonic states @xmath68 , @xmath69 and sections @xmath70 that satisfy @xmath71 .
the technique in @xcite allows to prove the following well posedness result .
[ 1thm : cauchyproblem ] assume that @xmath24 satisfies conditions ( * @xmath580*)-(*@xmath582 * ) .
for every @xmath72 and @xmath73 such that @xmath74\neq 0\ ] ] there exist positive @xmath75 , @xmath76 such that for all @xmath70 with @xmath77 there exists a semigroup @xmath78 with the following properties : 1 .
2 . for all @xmath80 , @xmath81 and for all @xmath82 , @xmath83 .
3 . for all @xmath84 and for all @xmath85 , @xmath86 4 . if @xmath80 is piecewise constant , then for @xmath87 small , @xmath88 is the gluing of solutions to riemann problems at the points of jump in @xmath52 and at the junction at @xmath89 .
5 . for all @xmath90 , the orbit @xmath91 is a @xmath24-solution to ( [ 1eq : e ] ) with initial datum @xmath50 .
the proof is postponed to section [ 1sec : tech ] . above @xmath92 , with @xmath93 , are the right eigenvectors of @xmath94 , see ( [ 1eigenv ] ) .
moreover , by _ solution to the riemann problems at the points of jump _ we mean the usual lax solution , see ( * ? ? ?
* chapter 5 ) , whereas for the definition of _ solution to the riemann problems at the junction _ we refer to ( * ? ? ?
* definition 2.1 ) . the same procedure used in ( * ? ? ?
* paragraph 2.2 ) allows now to construct the semigroup generated by ( [ 1eq : e ] ) in the case of a pipe with piecewise constant section @xmath95-\infty , x_1 ] } + \sum_{j=1}^{n-1 } a_j \ , \chi_{[x_{j } , x_{j+1 } [ } + a_{n } \ , \chi_{[x_n , + \infty[}\ ] ] with @xmath96 . in each segment @xmath97x_j , x_{j+1 } \right[$ ] , the fluid is modeled by ( [ 1eq : e ] ) . at each junction @xmath98 , we require condition ( [ 1eq : junction ] ) ,
namely @xmath99 we omit the formal definition of @xmath24-solution to ( [ 1eq : e])([1eq : junction ] ) in the present case , since it is an obvious iteration of definition [ 1def : weaksolution ] .
the natural extension of theorem [ 1thm : cauchyproblem ] to the case of ( [ 1eq : e])([1eq : psii ] ) is the following result .
[ 1thm : nn ] assume that @xmath24 satisfies conditions ( * @xmath580*)-(*@xmath582 * ) .
for any @xmath100 and any @xmath73 , there exist positive @xmath101 such that for any pipe s profile satisfying @xmath102\bar a-
\delta , \bar a+\delta\right[\right ) \mbox { with } { \mathrm{tv}}(a ) < m\ ] ] there exists a piecewise constant stationary solution @xmath103-\infty , x_1\right [ } + \sum_{j=1}^{n-1 } \hat u_j \chi_{\left]x_{j } , x_{j+1}\right [ } + \hat u_n \chi_{\left]x_{n } , + \infty\right[}\ ] ] to ( [ 1eq : e])([1eq : psii ] ) satisfying @xmath104 and a semigroup @xmath105 such that 1 .
@xmath107 is the identity and for all @xmath108 , @xmath109 .
3 . for all @xmath110 and for all @xmath111 ,
@xmath112 4 .
if @xmath113 is piecewise constant , then for @xmath87 small , @xmath88 is the gluing of solutions to riemann problems at the points of jump in @xmath52 and at each junction @xmath98 .
5 . for all @xmath113 , the orbit
@xmath114 is a weak @xmath24-solution to ( [ 1eq : e])([1eq : psii ] ) .
we omit the proof , since it is based on the natural extension to the present @xmath26 case of ( * ? ? ?
* theorem 2.4 ) .
remark that , as in that case , @xmath75 and @xmath76 depend on @xmath115 only through @xmath116 and @xmath30 . in particular ,
all the construction above is independent from the number of points of jump in @xmath115 . in this paragraph , the pipe s section @xmath115 is assumed to satisfy @xmath117\bar a-\delta , \bar a+\delta\right [ \right ) \mbox { for suitable } \delta > 0 , \
\bar a > \delta \\ { \mathrm{tv}}(a ) < m \mbox { for a suitable } m > 0 \\
a'(x ) = 0 \mbox { for a.e . } x \in { { \mathbb{r}}}\setminus [ -x , x ] \mbox { for a suitable } x > 0 \ , .
\end{array } \right . \!\!\ ] ] the same procedure used in ( * ? ? ?
* theorem 2.8 ) allows to construct the semigroup generated by ( [ 1eq : e1 ] ) in the case of a pipe which satisfies ( [ 1eq : a1 ] ) .
indeed , thanks to theorem [ 1thm : nn ] , we approximate @xmath115 with a piecewise constant function @xmath118 .
the corresponding problems to ( [ 1eq : e])([1eq : psii ] ) generate semigroups @xmath119 defined on domains characterized by uniform bounds on the total variation and that are uniformly lipschitz in time . here
, uniform means also independent from the number of junctions .
therefore , we prove the pointwise convergence of the @xmath119 to a limit semigroup @xmath120 , along the same lines in ( * ? ? ?
* theorem 2.8 ) .
[ 1sec : coupl cond ] this section is devoted to different specific choices of ( [ 1eq : psi ] ) .
[ [ s - solutions ] ] * ( s)-solutions * + + + + + + + + + + + + + + + we consider first the coupling condition inherited from the smooth case . for smooth solutions and pipes sections , system ( [ 1eq : e1 ] )
is equivalent to the @xmath26 balance law @xmath121 the stationary solutions to ( [ 1eq : e1 ] ) are characterized as solutions to @xmath122 as in the @xmath56 case of the @xmath25-system , the smoothness of the sections induces a unique choice for condition ( [ 1eq : psi ] ) , see ( * ? ? ?
* ( 2.3 ) and ( 2.19 ) ) , which reads @xmath123\ ] ] where @xmath124 is a smooth monotone function satisfying @xmath125 and @xmath126 , for a suitable @xmath127 .
@xmath128 are the @xmath3 and @xmath7 component in the solution to ( [ 1eq : stationary ] ) with initial datum @xmath129 assigned at @xmath130 .
note that , by the particular form of ( [ 1eq : psismooth ] ) , the function @xmath24 is independent both from the choice of @xmath131 and from that of the map @xmath115 , see ( * ? ? ?
* in proposition 2.7 ) .
[ [ p - solutions ] ] * ( p)-solutions * + + + + + + + + + + + + + + + the particular choice of the coupling condition in ( * ? ? ?
* section 3 ) can be recovered in the present setting .
indeed , conditions * ( m ) * , * ( e ) * and * ( p ) * therein amount to the choice @xmath132,\ ] ] where @xmath133 and @xmath134 are the pipe s sections .
consider fluid flowing in a horizontal pipe with an elbow or kink , see @xcite .
then , it is natural to assume the conservation of the total linear momentum along directions dependent upon the geometry of the elbow .
as the angle of the elbow vanishes , one obtains the condition above , see ( * ? ? ?
* proposition 2.6 ) .
[ [ l - solutions ] ] * ( l)-solutions * + + + + + + + + + + + + + + + we can extend the construction in @xcite to the @xmath26 case ( [ 1eq : e ] ) . indeed , the conservation of the mass and linear momentum in @xcite with the conservation of the total energy for the third component lead to the choice @xmath135,\ ] ] where @xmath133 and @xmath134 are the pipe s sections .
the above is the most immediate extension of the standard definition of lax solution to the case of the riemann problem at a junction . [
[ p - solutions-1 ] ] * ( p)-solutions * + + + + + + + + + + + + + + + following @xcite , motivated by the what happens at the hydrostatic equilibrium , we consider a coupling condition with the conservation of the pressure @xmath136 in the second component of @xmath24
. thus @xmath137,\ ] ] where @xmath133 and @xmath134 are the pipe s sections .
[ 1prop : otherpsi ] for every @xmath100 and @xmath138 , each of the coupling conditions @xmath24 in ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) , ( [ 1eq : psip ] ) satisfies the requirements ( * @xmath580*)-(*@xmath582 * ) and ( [ 1eq : det ] ) . in the case of ( [ 1eq : psip ] )
, we also require that the fluid is perfect , i.e. that ( [ 1eq : perfect ] ) holds .
the proof is postponed to section [ 1sec : tech ] .
thus , theorem [ 1thm : cauchyproblem ] applies , yielding the well posedness of ( [ 1eq : e])([1eq : junction ] ) with each of the particular choices of @xmath24 in ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) , ( [ 1eq : psip ] ) .
[ 1sec : blow - up ] in the previous results a key role is played by the bound on the total variation @xmath30 of the pipe s section .
this requirement is intrinsic to problem ( [ 1eq : e])([1eq : junction ] ) and not due to the technique adopted above .
indeed , we show below that in each of the cases ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) , ( [ 1eq : psip ] ) , it is possible to choose an initial datum and a section @xmath139)$ ] with @xmath140 arbitrarily small , such that the total variation of the corresponding solution to ( [ 1eq : e])([1eq : junction ] ) becomes arbitrarily large .
consider the case in figure [ 1fig : f ] .
a wave @xmath141 hits a junction where the pipe s section increases by @xmath142 . from this interaction ,
the wave @xmath143 arises , which hits a second junction , where the pipe section decreases by @xmath142.,width=226 ] a wave @xmath141 hits a junction where the pipe s section increases by , say , @xmath144 .
the fastest wave arising from this interaction is @xmath143 , which hits the second junction where the section diminishes by @xmath142 . solving the riemann problem at the first interaction amounts to solve the system @xmath145 where @xmath146 , see figure [ 1fig : notinter ] for the definitions of the waves strengths @xmath147 and @xmath141 . above
, @xmath148 is the map defined in ( [ 1eq : t ] ) , which in turn depends from the specific condition ( [ 1eq : psi ] ) chosen . in the expansions below , we use the @xmath149 variables , thus setting @xmath150 throughout this section . ) and ( [ 1second]).,width=151 ] differently from the case of the @xmath56 @xmath25-system in @xcite , here we need to consider the second order expansion in @xmath151 of the map @xmath148 ; that is @xmath152 the explicit expressions of @xmath153 and @xmath154 in ( [ 1eq : tt ] ) , for each of the coupling conditions ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) , ( [ 1eq : psip ] ) , are in section [ 1subs : conti ] . inserting ( [ 1eq : tt ] ) in the first order expansions in the wave s sizes of ( [ 1eq : pbr ] ) , with @xmath155 for @xmath93 as in ( [ 1tangent vectors ] ) , we get a linear system in @xmath156 . now , introduce the fluid speed @xmath157 and the adimensional parameter @xmath158 a sort of _ `` mach number''_. obviously , @xmath159 $ ] for @xmath146 .
we thus obtain an expression for @xmath160 of the form @xmath161 the explicit expressions of @xmath162 and @xmath163 in ( [ 1first ] ) are in section [ 1subs : conti ] .
remark that the present situation is different from that of the @xmath56 @xmath25-system considered in @xcite . indeed , for the @xmath25-system @xmath164 , while here it is necessary to compute the second order term in @xmath165 .
concerning the second junction , similarly , we introduce the parameter @xmath166 which corresponds to the state @xmath167 .
recall that @xmath69 is defined by @xmath168 , see figure [ 1fig : notinter ] and section [ 1subs : conti ] for the explicit expressions of @xmath169 .
we thus obtain the estimate @xmath170 where @xmath171 .
now , at the second order in @xmath165 and at the first order in @xmath141 , ( [ 1first ] ) and ( [ 1second ] ) give @xmath172 indeed , computations show that @xmath173 vanishes at the first order in @xmath165 , as in the case of the @xmath25-system .
the explicit expressions of @xmath174 are in section [ 1subs : conti ] .
it is now sufficient to compute the sign of @xmath174 .
if it is positive , then repeating the interaction in figure [ 1fig : f ] a sufficient number of times leads to an arbitrarily high value of the refracted wave @xmath175 and , hence , of the total variation of the solution @xmath52 .
below , section [ 1sec : tech ] is devoted to the computations of @xmath174 in the different cases ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) and ( [ 1eq : psip ] ) . to reduce the formal complexities of the explicit computations below
, we consider the standard case of an ideal gas characterized by ( [ 1eq : perfect ] ) with @xmath176 .
the results of these computations are in figure [ 1fig : chi ] .
they show that in all the conditions ( [ 1eq : junction ] ) considered , there exists a state @xmath146 such that @xmath177 , showing the necessity of condition ( [ 1eq : a0 ] ) .
however , in case * ( l ) * , it turns out that @xmath174 is negative on an non trivial interval of values of @xmath178 . as a function of @xmath178 .
top , left , case * ( s ) * ; right , case * ( p ) * ; bottom , left , case * ( l ) * ; right , case * ( p)*. note that in all four cases , @xmath174 attains strictly positive values , showing the necessity of the requirement ( [ 1eq : a0]).,title="fig : " ] as a function of @xmath178 .
top , left , case * ( s ) * ; right , case * ( p ) * ; bottom , left , case * ( l ) * ; right , case * ( p)*. note that in all four cases , @xmath174 attains strictly positive values , showing the necessity of the requirement ( [ 1eq : a0]).,title="fig : " ] + as a function of @xmath178 .
top , left , case * ( s ) * ; right , case * ( p ) * ; bottom , left , case * ( l ) * ; right , case * ( p)*. note that in all four cases , @xmath174 attains strictly positive values , showing the necessity of the requirement ( [ 1eq : a0]).,title="fig : " ] as a function of @xmath178 .
top , left , case * ( s ) * ; right , case * ( p ) * ; bottom , left , case * ( l ) * ; right , case * ( p)*. note that in all four cases , @xmath174 attains strictly positive values , showing the necessity of the requirement ( [ 1eq : a0]).,title="fig : " ] if @xmath179 is chosen in this interval , the wave @xmath175 in the construction above is not magnified by the consecutive interactions .
the computations leading to the diagrams in figure [ 1fig : chi ] are deferred to section [ 1subs : conti ] .
[ 1sec : tech ] we recall here basic properties of the euler equations ( [ 1eq : e1 ] ) , ( [ 1eq : e ] ) .
the characteristic speeds and the right eigenvectors have the expressions @xmath180
r_{1 } = \left [ \!\!\ !
\begin{array}{c } -\rho \\ \rho c - q \\ qc - e - p \end{array } \!\!\
! \right ] & r_{2 } = \left [ \!\!\ !
\begin{array}{c } \rho \\ q \\ e+p-\frac{\rho^{2}c^2}{\partial_{e}p } \end{array } \!\!\ !
\right ] & r_{3 } = \left [ \!\!\ !
\begin{array}{c } \rho
\\ q+\rho c \\
e+p+qc \end{array } \!\!\ !
\right ] \end{array}\ ] ] whose directions are chosen so that @xmath181 for @xmath93 . in the case of an ideal gas ,
the sound speed @xmath182 becomes @xmath183 the shock and rarefaction curves curves of the first and third family are : @xmath184^{-1}\,dp } \quad\\ s(\rho , e)=s(\rho_o , e_o ) \end{array } \right .
\mbox { for } \begin{array}{rcl } \sigma & \geq & 0 \\
\rho & \leq & \rho_o \\ v & \geq & v_o \\ e & \leq & e_o \end{array } \\
r_3(u_o,\sigma ) & = & \left\ { \begin{array}{l } \rho=\sigma+\rho_o \\ v = v_o+\displaystyle{\int_{p_o}^p [ ( \rho\,c)(p , s_o)]^{-1}\,dp } \quad\\ s(\rho , e)=s(\rho_o , e_o ) \end{array } \right .
\mbox { for } \begin{array}{rcl } \sigma & \geq & 0 \\
\rho & \geq & \rho_o \\ v & \geq & v_o \\ e & \geq & e_o \end{array}\end{aligned}\ ] ] the 1,2,3-lax curves have the expressions @xmath185 their reversed counterparts are @xmath186^{-1}\,dp } \quad\\ s(\rho , e)=s(\rho_o , e_o ) \end{array } \right .
\mbox { for } \begin{array}{rcl } \sigma & \geq & 0 \\
\rho & \geq & \rho_o \\
v & \leq & v_o \\ e & \geq & e_o \end{array } \\
r_3^{-}(u_o,\sigma ) & = & \left\ { \begin{array}{l } \rho=-\sigma+\rho_o \\ v = v_o+\displaystyle{\int_{p_o}^p [ ( \rho\,c)(p , s_o)]^{-1}\,dp } \quad \\ s(\rho , e)=s(\rho_o , e_o ) \end{array } \right .
\mbox { for } \begin{array}{rcl } \sigma & \geq & 0 \\
\rho & \leq & \rho_o \\
v & \leq & v_o \\ e & \leq & e_o \end{array}\end{aligned}\ ] ] and @xmath187 in the @xmath188 space , for a perfect ideal gas , the tangent vectors to the lax curves are : @xmath189 \,,\
\tilde r_2 = \left[\!\ !
\begin{array}{c } 1 \\ \frac{q}{\rho } \\
-\frac{e}{\rho } \end{array } \!\!\right ] \,,\ \tilde r_3 = \left[\!\ !
\begin{array}{c } 1 \\ \frac{q}{\rho } -\sqrt{\gamma ( \gamma-1 ) e } \\ ( \gamma-1 ) \frac{e}{\rho } \end{array } \!\!\right].\;\end{aligned}\ ] ] [ 1prop : lax curves ] let @xmath190 be the @xmath191-th lax curve and @xmath192 be the reversed @xmath191-th lax curve through @xmath193 , for @xmath93 .
the following equalities hold : @xmath194 @xmath195 @xmath196 @xmath197 @xmath198 theorem [ 1thm : cauchyproblem ] following ( * ? ? ?
* proposition 4.2 ) , the @xmath26 system ( [ 1eq : e ] ) defined for @xmath199 can be rewritten as the following @xmath200 system defined for @xmath201 : @xmath202 the relations between @xmath203 and @xmath34 , between @xmath204 and the flow in ( [ 1eq : e ] ) being @xmath205 \quad\mbox { and } \quad \mathcal{f}(u ) = \left [ \begin{array}{c } u_2 \\
p(u_1,u_2,u_3 ) \\ f(u_1,u_2,u_3 ) \\ u_5 \\
p(u_4,u_5,u_6 ) \\
f(u_4,u_5,u_6 ) \end{array } \right]\ ] ] with @xmath206 and @xmath207 defined in ( [ 1eq : epf ] ) ; whereas the boundary condition in ( [ 1eq : uf ] ) is related to ( [ 1eq : junction ] ) by @xmath208 for fixed sections @xmath48 and @xmath49 .
the thesis now follows from ( * ? ? ?
* theorem 2.2 ) .
indeed , the assumptions * ( @xmath209 ) * , * ( b ) * and * ( f ) * therein are here satisfied .
more precisely , condition * ( @xmath209 ) * follows from the choice ( [ 1subsonic ] ) of the subsonic region @xmath210 .
simple computations show that condition * ( b ) * reduces to @xmath211 \\
= \det \left [ d_{u^{-}}\psi\cdot r_{1}(\bar u ) \quad -d_{u^{+}}\psi \cdot r_{2}(\bar u ) \quad -d_{u^{+}}\psi \cdot r_{3}(\bar u ) \right ] \\
= \det \left [ d_{u^{-}}\psi\cdot r_{1}(\bar u ) \quad d_{u^{+}}\psi \cdot r_{2}(\bar u ) \quad d_{u^{+}}\psi \cdot r_{3}(\bar
u ) \right ] , \end{array}\ ] ] which is non zero for assumption if @xmath138 and @xmath212 . condition * ( f ) * needs more care .
indeed , system ( [ 1eq : uf ] ) is _ not _ hyperbolic , for it is obtained gluing two copies of the euler equations ( [ 1eq : e ] ) .
nevertheless , the two systems are coupled only through the boundary condition , hence the whole wave front tracking procedure in the proof of ( * ? ? ?
* theorem 2.2 ) applies , see also ( * ? ? ?
* proposition 4.5 ) . proposition [ 1prop : otherpsi ] it is immediate to check that each of the coupling conditions ( [ 1eq : psismooth ] ) , ( [ 1eq : psicri ] ) , ( [ 1eq : psil ] ) , ( [ 1eq : psip ] ) satisfies the requirements ( * @xmath580 * ) and ( * @xmath581 * ) . to prove that ( * @xmath582 * ) is satisfied , we use an _ ad hoc
argument for condition * ( s)*. in all the other cases , note that the function @xmath24 admits the representation @xmath213 .
therefore , ( * @xmath582 * ) trivially holds .
we prove below ( [ 1eq : det ] ) in each case separately .
note however that for any of the considered choices of @xmath24 , @xmath214 so that ( [ 1eq : det ] ) reduces to @xmath215 \\ & = & -\det d_{u^{+}}\psi \cdot \det \left [ r_{1}(\bar
u ) \quad r_{2}(\bar u ) \quad r_{3}(\bar u ) \right ] \ , .
\end{aligned}\ ] ] thus , it is sufficient to prove that @xmath216 . to prove that the coupling condition ( [ 1eq : psismooth ] ) satisfies ( * @xmath582 * ) , simply use the additivity of the integral and the uniqueness of the solution to the cauchy problem for the ordinary differential equation ( [ 1eq : stationary ] ) .
next , we have @xmath217 since @xmath218 for all @xmath219 , because @xmath220 . thus ,
@xmath24 in ( [ 1eq : psismooth ] ) satisfies @xmath221 \!\ ! \\
= -\bar a^{3 } \ , \lambda_{1}(\bar u ) \ , \lambda_{2}(\bar u ) \ , \lambda_{3}(\bar u ) , \end{array}\ ] ] which is non zero if @xmath138 . concerning condition ( [ 1eq : psicri ] )
, we have @xmath222 \!\ !
\\ = -\bar a^{2}\lambda_{1}(\bar u)\lambda_{2}(\bar u)\lambda_{3}(\bar u ) , \end{array}\ ] ] which is non zero if @xmath138 . finally , concerning condition ( [ 1eq : psip ] ) , @xmath224 \!\ ! \\
= \bar a^{2}\lambda_{2}(\bar u)\left ( c^{2}+\lambda_{2}^{2}(\bar u)\frac{{\partial_e \bar p}}{\bar \rho}\right ) , \end{array}\ ] ] which is non zero if @xmath138 and if the fluid is perfect , i.e. ( [ 1eq : perfect ] ) holds .
let @xmath24 be defined in ( [ 1eq : psismooth ] ) and set @xmath225 where the functions @xmath226 have the same meaning as in ( [ 1eq : psismooth ] ) .
a perturbative method allows to compute the solution to ( [ 1eq : stationary ] ) with a second order accuracy in @xmath165 .
then , long elementary computations allow to get explicitly the terms @xmath153 and @xmath154 in ( [ 1eq : tt ] ) of the second order expansion of @xmath148 : @xmath227 \\
g(\rho , q , e ) & = & \left [ \begin{array}{l } \displaystyle -{{4\left({\vartheta}^3 - 2\,{\vartheta}^2\right ) } \over{3\left ( { \vartheta}^3 - 3\,{\vartheta}^2 + 3\,{\vartheta}-1\right ) } } \ ; \rho \\
\\ \displaystyle -{{70\,{\vartheta}^4 - 257\,{\vartheta}^3 + 342\,{\vartheta}^2 - 207\,{\vartheta}+36}\over{18\left ( { \vartheta}^3 - 3\,{\vartheta}^2 + 3\,{\vartheta}-1\right ) } } \ ; e \end{array } \right]\,.\end{aligned}\ ] ] moreover , the coefficients @xmath162 , @xmath163 in ( [ 1first ] ) read @xmath228 next , @xmath174 is given by @xmath229 let @xmath24 be defined in ( [ 1eq : psicri ] ) .
with reference to ( [ 1eq : tt ] ) , we show below explicitly the terms @xmath153 and @xmath154 in ( [ 1eq : tt ] ) of the second order expansion of @xmath148 , @xmath230 \\
g(\rho , q , e ) & = & \left [ \begin{array}{l } \displaystyle { { 64\left ( { \vartheta}^3 + 3\,{\vartheta}^2\right ) } \over{27\left ( { \vartheta}^3 - 3\ , { \vartheta}^2 + 3\,{\vartheta}-1\right ) } } \ ; \rho \\ q \\
\displaystyle -{{565\,{\vartheta}^4 - 1599\,{\vartheta}^3 + 927\,{\vartheta}^2 - 405\ , { \vartheta}}\over{81\left ( { \vartheta}^3 - 3\,{\vartheta}^2 + 3\,{\vartheta}-1\right ) } } \ ; e \end{array } \right]\,.\end{aligned}\ ] ] moreover , the coefficients @xmath162 , @xmath163 in ( [ 1first ] ) read @xmath231 next , @xmath174 is given by @xmath232 let @xmath24 be defined in ( [ 1eq : psil ] ) .
then , @xmath233 \qquad \mbox { and } \qquad g(\rho , q , e ) = \left [ \begin{array}{l } \displaystyle -{{4\,{\vartheta}}\over{3\left ( { \vartheta}-1\right ) } } \ ; \rho
\\ q \\
\displaystyle -{{35\,{\vartheta}^2 - 9\left ( 4\,{\vartheta}-1\right ) } \over{9\left ( { \vartheta}-1\right ) } } \ ; e \end{array } \right].\ ] ] the coefficients @xmath162 , @xmath163 in ( [ 1first ] ) read @xmath234 so that @xmath174 is @xmath235 let @xmath24 be defined in ( [ 1eq : psip ] ) . with reference to ( [ 1eq : tt ] ) , @xmath236 \\
g(\rho , q , e ) & = & \left [ \begin{array}{l } \displaystyle -{{4\left ( { \vartheta}^3 + 3\,{\vartheta}^2\right ) } \over{4\left ( 2\,{\vartheta}^3 + 9\,{\vartheta}^2\right ) + 27\left ( 2\,{\vartheta}+1\right ) } } \ ; \rho \\ q \\ \displaystyle { { 12\left ( { \vartheta}^3 + 2\,{\vartheta}^2\right ) } \over{4\left ( 2\,{\vartheta}^3 + 9\,{\vartheta}^2\right ) + 27\left ( 2\,{\vartheta}+1\right ) } } \ ; e \end{array } \right],\end{aligned}\ ] ] with @xmath162 and @xmath163 given by @xmath237 so that @xmath238 | this paper is devoted to the extension to the full @xmath0 euler system of the basic analytical properties of the equations governing a fluid flowing in a duct with varying section .
first , we consider the cauchy problem for a pipeline consisting of 2 ducts joined at a junction .
then , this result is extended to more complex pipes .
a key assumption in these theorems is the boundedness of the total variation of the pipe s section .
we provide explicit examples to show that this bound is necessary . _
keywords : _ conservation laws at junctions , coupling conditions at junctions . _
2000 msc : _
35l65 , 76n10 |
the currently accepted scenario for canonical glasses includes different relaxation mechanisms that are universally present in all systems .
experimentally , these different mechanisms are clearly revealed by the dielectric spectra that shows a broad low frequency peak referred to as @xmath3-relaxation @xcite and a higher frequency peak or shoulder usually called johari - goldstein @xmath4-relaxation @xcite .
the @xmath3-relaxation is generally attributed to processes involving the cooperative dynamics of regions of molecules @xcite .
the microscopical origin of the @xmath4 relaxation is still a matter of wide debate@xcite .
the proposed models explain this peak as a consequence of the nonuniformity of the glassy state involving only local regions in which molecules can diffuse ( islands of mobility ) .
an alternative homogeneous explanation attributes the secondary relaxation phenomena to small - angle reorientations of all the molecules @xcite .
systems having translational crystalline order but rotational degrees of freedom also display a glassy behavior .
since in these cases one type of degree of freedom is completely absent , they represent simplified models on which to test theoretical concepts on glassy dynamics @xcite . compounds of molecules of the type cbr@xmath1cl@xmath2 , with n=0,1,2 , are examples of systems with this characteristics @xcite .
they have a series of solid - solid phase transitions attributed to the ability of the molecules to acquire rotational degrees of freedom as the temperature is increased .
all these compounds crystallize from the melting to a plastic fcc phase
. further reduction of the temperature leads to a @xmath6 monoclinic phase @xcite . the cases n=1 and n=2 exhibit a glass transition at 90 k that is clearly visible through calorimetric techniques and its structure shows disorder in the position of the cl and br atoms @xcite .
the case n=0 can not display such a disorder and , in fact , the calorimetric curve does not show a glass transition at low temperatures @xcite .
the dielectric spectra of cbrcl@xmath7 and cbr@xmath8cl@xmath8 was reported by zuriaga _
@xcite in the temperature range 100 250 k and 100 210 k , respectively . at the lower end of the temperature range
the most relevant characteristic is that both spectra display a well - defined shoulder on the high frequencies flank of the @xmath3-peak , which is attributed to the @xmath4-relaxation .
since ccl@xmath0 has no molecular dipole moment it is not accessible to dielectric experiments but it can be studied using nuclear quadrupolar resonance .
interestingly , the resolution of the nqr spectra for ccl@xmath0 is well superior to the corresponding spectra for cbr@xmath8cl@xmath8 and cbrcl@xmath7 and the two techniques complement each other . on the other hand , the nqr experiments are limited to a temperature range between 77 k and 140 k , with the upper end determined by the broadening of the signal . the picture that emerges from the combined analysis
is that the three compounds have a very similar dynamic evolution in the monoclinic phase as a function of temperature @xcite .
the analysis of the isostructural ccl@xmath0 shows that nonequivalent molecules in the unit cell perform reorientational jumps at different time scales due to their different crystalline environments .
these results support the conclusion that the dynamic heterogeneity is intimately related to the secondary relaxation observed in these compounds @xcite .
this work is a direct extension of our previous works involving halogenomethanes @xcite . in this case
we have studied the monoclinic phase of ccl@xmath0 using extensive molecular dynamics simulations on a fully anisotropic cell , and the results were compared with an analytical stochastic model .
the advantage of this particular system is that it allows for the determination of rotational correlation times from nqr measurements with sufficient precision to distinguish between the nonequivalent groups of molecules .
the same distinction is easily done in the molecular dynamics simulations , which in turn allows for an analysis at the individual nonequivalent groups and correlate the results with the location of the molecules in the lattice .
the temperature range affordable by the simulations covers from the fcc phase down to 160 k. simulations at lower temperatures became impossible as the relaxation times go beyond 10 @xmath9s .
is is found that , when the spin - lattice relaxation times from the experiments and from the simulations are plotted in the 100 220 k temperature range , the two sets of curves correspond to each other very well for the five distinguishable modes suggesting that ccl@xmath0 behaves as a strong glass in the whole temperature range , in spite of the fact that the compound can not glassify due to its lack of orientational entropy .
ccl@xmath0 is a tetrahedral molecule having three equivalent @xmath10 and four equivalent @xmath11 symmetry axes .
following our previous work@xcite , we have modeled the ccl@xmath0 molecule as a rigid , non - polarizable tetrahedron with the carbon atom at its center and a chlorine atom at each one of the vertices .
the interaction between molecules is represented by a combination of lennard - jones and coulombic terms summarized in table [ table : potential ] .
the cross interaction between atoms of different type is calculated by applying the lorentz - berthelot combination rules , i.e. , geometrical mean for @xmath12 and arithmetic mean for @xmath13 . a spherical cut - off at 1.5 nm
was imposed on all intermolecular interactions .
periodic boundary conditions were imposed in all three cartesian directions .
l c c c c c + & @xmath12 [ kj / mol ] & @xmath13 [ nm ] & @xmath14 [ e ] & + c & 0.22761 & 0.37739 & -0.696 & c - cl & 0.1766 + cl & 1.09453 & 0.34667 & 0.174 & cl - cl & 0.2884 + the monoclinic crystal structure of ccl@xmath0 , resolved by cohen et _ al .
_ at 195 k @xcite , corresponds to the @xmath6 space group .
the unit cell , which contains @xmath15=32 molecules , has the following lattice parameters : @xmath16=2.0181 nm , @xmath17=1.1350 nm , @xmath18=1.9761 nm and angle @xmath19 . using the experimental crystalline structure as initial coordinates , we constructed a simulation supercell containing 512 molecules , which correspond to 16 monoclinic unit cells .
this supercell was prepared by replicating the experimental unit cell twice on the @xmath20 and @xmath21 directions and four times in the @xmath22 direction .
the molecular dynamics simulations , conducted under npt conditions , have been carried out using the gromacs v5.0.2 simulation package .
atom - atom distances within each molecule were kept constants with the shake algorithm .
the simulations were started with a 10 ns equilibration run with the pressure controlled by a berendsen barostat and the temperature controlled with the v - rescale thermostat .
the berendsen weak coupling method ensured a smooth approach to equilibrium with no disruptions to the simulation cell .
the classical newton s equations were integrated using the leap - frog algorithm and the time step of the integration of the equations of motion was set to 1 fs .
the production runs were extended up to 100 ns or 10 @xmath9s depending on the temperature , using in this case a time step of 5 fs .
the control of the temperature was made by using a nos - hoover thermostat , with a time constant of 2.0 ps .
the pressure was maintained constant by using a fully anisotropic parrinello - rahman barostat with a reference pressure of 1 atm .
the study covered temperatures ranging from 160 k to 230 k , in steps of 10 k. the monoclinic structure was stable in the whole analyzed temperature range ( see supporting information file ) .
the spatial positions of the molecules in the unit cell of the system are defined through the application of 8 symmetry operations over the 4 nonequivalent molecules ( 16 cl atoms with nonequivalent positions ) . as a consequence , the system has four distinctive groups of molecules that we will refer to as g@xmath23 , g@xmath24 , g@xmath25 and g@xmath26 . in our simulation cell ,
each group contains 128 molecules .
each molecule belonging to a given group has the same specific arrangement of neighboring molecules .
all the molecular reorientation processes occurring during the simulations are sudden large - angle jumps of the cl atoms .
the jumps were detected using a running test algorithm based on a signal to noise measure and adapted from the work of carter and cross@xcite . in this method ,
the molecular jumps are detected by a spike in the test function as explained in the supporting information file .
after analyzing the trajectories for all the temperatures , the time @xmath27 of every single jump was registered .
the angle described by a c - cl bond upon a jump was calculated using the average direction of the bond before and after the jump .
namely , defining @xmath28 as the bond between the central c atom and the @xmath29-th cl atom of the same molecule , the angle jump at time @xmath27 is defined as the angle between @xmath30 and @xmath31 . here , the angular brackets represent time average and the subscript @xmath32 indicates the lapse between @xmath33 and @xmath34 . in figure [ angles_160 ]
we show the relative frequency of the reorientation angles calculated from the simulation at 160 k and for the four different groups of molecules .
all the curves peak at the tetrahedral angle , which is the angle that should be observed upon @xmath5 type rotations . indeed , a careful inspection confirmed that for all temperatures there were just a handful of cases corresponding to @xmath35 type rotations and therefore there were neglected in the analysis .
normalized frequency distribution of reorientation angles for the four distinct group of molecules , as indicated .
the results correspond to the production run at 160 k. the maximum at the tetrahedral angle indicate that all jumps are @xmath5 type rotations.,scaledwidth=40.0% ] in order to characterize the molecular rotations using standard self correlation methods , we calculate the relaxation of the bond orientation through self correlation functions defined as @xmath36 where @xmath37 represents a particular c - cl bond , @xmath38 represents the position of the c - cl@xmath39 bond at the initial condition , and the angular brackets represent the average over the 128 molecules of the same group .
the resulting self correlation functions are shown in figure [ cgroups ] with solid lines .
the correlation times for the reorientations of the molecules depend not only on the groups but also the curves reveal that one bond ( green lines ) maintains its orientation for a time considerably longer than the other three bonds .
this indicates that the molecules have preferential axes of rotation .
g@xmath25 exhibit the strongest rotational anisotropy , followed by g@xmath24 and g@xmath23 and only a minor effect is observed in g@xmath26 .
the strength of this anisotropic character decreases with increasing temperature as all the free energy barriers become relatively smaller . during a rotation of the type @xmath5 one of the c - cl bonds
remain in its place and therefore one cl atom does not change its average position .
the remaining three cl atoms undergo a jump of 109.5 degrees . each molecule has four possible axis for a @xmath5 type rotation that coincide with the c - cl molecular bonds .
as all the molecular bonds remain lying around the four initial orientations in the crystal it is possible to define , for each molecule , four axis of rotation relative to the overall orientation of the crystal . in this way we define the axes @xmath40 , for each one of the molecules in the simulation cell .
the frequencies of rotational jumps for the four groups of molecules along the four possible axes of rotation , averaged over each group , are reproduced in table [ table : w ] .
the numerical values of the frequencies span over four order of magnitude implying the existence of different dynamical modes in the crystal as already suggested by the rotational self correlation functions .
c|cccc|cccc + & & + axis & g@xmath23 & g@xmath24 & g@xmath25 & g@xmath26 & g@xmath23 & g@xmath24 & g@xmath25 & g@xmath26 + @xmath41 & @xmath42 & @xmath43 & @xmath44 & @xmath43 & @xmath45 & @xmath46 & @xmath47 & @xmath48 + @xmath49 & @xmath50 & @xmath51 & @xmath43 & @xmath52 & @xmath53 & @xmath54 & @xmath55 & @xmath56 + @xmath57 & @xmath58 & @xmath59 & @xmath60 & @xmath61 & @xmath62 & @xmath63 & @xmath64 & @xmath65 + @xmath66 & @xmath67 & @xmath68 & @xmath43 & @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath73 + auto - correlation functions @xmath74 of the 4 bonds c - cl@xmath39 averaged over molecules within the four groups of the system .
the solid lines correspond to the md simulation results and the dashed lines are derived from the master equation eq .
( [ master_equation ] ) .
different colors correspond to the four different bonds.,scaledwidth=40.0% ] in order to rationalize the behavior of the functions @xmath74 we developed a simple analytical model that derive these functions from the relative reorientation frequencies .
we propose a master equation that describes the probabilities of occupation of the four available sites in the molecule for the cl atoms .
let us define the stochastic variables @xmath75 that describe the position of the j - th cl atom on the four available sites .
then , each @xmath75 takes values over a range represented by @xmath76 , with a probability distribution @xmath77 over this range .
assuming that a reorientational jump in the molecule is a markovian process , the master equation that describes the probability @xmath77 of the state of @xmath75 is @xmath78 \ , ; j=1,\dots,4,\ ] ] where @xmath79 is the transition probability per unit time from the site @xmath80 to the site @xmath81 .
these equations may be written in matrix form as @xmath82 where @xmath83 is the vector @xmath84 , and @xmath85 is the @xmath86 matrix @xmath87
( [ dp ] ) is an homogeneous first order differential equation with constant coefficients .
the solution is @xmath88 where @xmath89 is the matrix that diagonalize @xmath85 , @xmath90 . assuming that c3 jumps are the exclusive mechanism of molecular reorientation , the @xmath79 can be approximated using the frequency of rotations @xmath91 around the axes @xmath92 .
these frequencies were calculated from the md simulations and are summarized in table [ table : w ] .
they can be interpreted as the probabilities @xmath91 for rotations per unit time around each one of the four possible axis .
frequencies smaller than 10@xmath93 ns@xmath94 correspond to less than 5 rotational jumps in 10 @xmath9s and therefore are negligible .
the tetrahedral constraint between the cl atoms imply that a @xmath95 transition can be achieved under two different single rotation events . considering all rotation events as independent processes
, we can calculate @xmath79 as the sum of the probability of the two possible events leading to a @xmath95 transition .
for example , @xmath96 , @xmath97 , @xmath98 and so on .
the factor of @xmath99 is due to the assumption that both directions of rotation are equally probably , which was tested to be statistically true . the next step is to solve the proposed master equation with the four different initial conditions , which are @xmath100 , @xmath101 , @xmath102 and @xmath103 .
let us denote the four corresponding solutions as @xmath104 , with @xmath105 .
then , the rotational auto correlation functions can be expressed in terms of the solution of the master equation as @xmath106.\ ] ] the resulting solutions @xmath74 are plotted on figure [ cgroups ] using dashed lines .
the agreement between the results directly measured from the simulation trajectories and those derived from the master equation is excellent , justifying the assumptions made in the derivation of the analytical model .
distributions of waiting times for two temperatures and the corresponding poisson curves ( in red ) for the fastest rotational axes of g@xmath23 , g@xmath24 and g@xmath25.,scaledwidth=40.0% ] the statistics behind the rotational jumps can be evaluated by studying the waiting time between successive jumps around the axes @xmath107 . on some of the axes
the rotational events are very rare and the length of our simulations do not allows to see a well defined pattern for those specific slow axes . nevertheless , for the fastest axes the process reveals itself as following a poisson distribution . in figure [ hist_160_220 ]
we show the distribution of waiting times for rotational events around the fastest axis of rotation for each group of molecules .
the histograms are directly obtained from the simulated trajectories .
the solid red lines in the plots for the three fastest groups correspond to @xmath108 , with the poisson parameter @xmath109 being the frequency of jumps around @xmath57 axes , @xmath110 from table [ table : w ] .
a similarity between the dynamical behavior of all the different groups could be inferred by looking at the self correlation functions @xmath74 and distribution of waiting times shown in figures [ cgroups ] and [ hist_160_220 ] , respectively .
in particular , g@xmath24 and g@xmath25 are the two groups with the closest quantitative behavior . in order to explore whether this kinetic resemblance has a correlation with the molecular arrangement we analyzed the orientation of the fastest axis for each one of the groups .
interestingly , the fastest axes of the molecules of g@xmath24 and g@xmath25 are parallel to each other in pairs that involve the first neighbors , as shown in figure [ parallel ] . for groups g@xmath23 and g@xmath26 there is a certain degree of spacial correlation , but it does not include the eight pairs of molecules of the unit cell . unit cell displaying the molecules of groups g@xmath23 , g@xmath24 , g@xmath25 and g@xmath26 in red , blue , violet and yellow respectively . the colored axis represent the fastest direction of rotation for each molecule.,scaledwidth=40.0% ] nqr spectroscopy allows to measure the spin - lattice relaxation time @xmath111 . since the low temperature crystalline structure of ccl@xmath0 has four nonequivalent molecules in the unit cell , the nqr spectral lines correspond to the 16 cl atoms of these four molecules .
then it is possible to measure the relaxation times @xmath111 , for each line , as function of temperature . as nuclear magnetization is proportional to the nuclear spin polarization ,
applying the slow reorientational model for @xmath111 by alexander and tzalmona it is possible to find the reorientation relaxation time as function of the jumps probabilities @xmath112 .
starting from the master equation for the nuclear polarization , zuriaga _ et al _ @xcite proposed a model for tetrahedral molecules that allow us to find the chlorine @xmath111 relaxation times as a function of reorientation frequencies @xmath91 around the four molecular axes . from the analytical model and the md simulations ,
16 relaxation times are obtained at each temperature , four for each group of molecules .
the results for @xmath111 vs. 1000/t are represented in figure [ taus ] along with the experimental results , which also include the cases of cbrcl@xmath7 and cbr@xmath8cl@xmath8 .
the existence of a preferential axis of rotation for three groups ( g@xmath23 , g@xmath24 and g@xmath25 ) allows the identification of three short times t@xmath113 of similar magnitude ( corresponding to the three cl atoms out of the fastest axis ) and a long time t@xmath114 ( for the cl atom on the fastest axis ) .
the remaining group ( g@xmath26 ) has no preferential axis of rotation and therefore the four relaxation times are all similar . the set of 16 relaxation times from the simulations are grouped in increasing order in the following five sets : 1 .
the shortest time correspond to the t@xmath113 describing g@xmath24 ( green triangles ) .
2 . the next time corresponds to t@xmath113 of g@xmath25 ( blue squares ) .
the third set is for t@xmath113 of g@xmath115 ( violet circles ) .
the fourth set includes t@xmath114 of g@xmath115 ( brown circles ) , t@xmath114 of g@xmath24 ( brown triangles ) and t@xmath116 of g@xmath26 ( brown diamonds ) .
the last set corresponds to t@xmath114 of g@xmath25 ( orange squares ) .
the black straight lines in figure [ taus ] are the result of a linear least - squares fit of the values @xmath117 vs. @xmath118 , including both , simulation and experimental results .
the are three main features that arise from figure 5 .
first , the simulations curves are in a good agreement with the experimental data , prolonging the same arrhenius curves from 100 to 230 k. second , the results for t@xmath116 spread along two order of magnitude in time , indicating the presence of fast and slow modes .
third , the overall behavior resembles @xmath3 and @xmath4 relaxation times of cbr@xmath119cl@xmath119 ( also shown in figure [ taus ] ) and cbrcl@xmath120 , showing that the dynamics of the non - glass former ccl@xmath0 has strong similarities with that of the other members of the isostructural series cbr@xmath1cl@xmath2 .
spin - lattice relaxation times obtained by molecular dynamics simulations ( full symbols ) . results from nqr studies of ccl@xmath0 @xcite ( empty symbols ) are shown .
red lines represents the behavior of @xmath121 ( upper ) and @xmath122 ( lower ) of cbrcl@xmath120 and cbr@xmath119cl@xmath119 taken from ref .
we performed extensive molecular dynamics simulations of ccl@xmath0 in the monoclinic cell . by studying time reorientational auto - correlation functions as a function of temperature we found distinct behaviors for the four different c - cl bonds in the molecule .
as a consequence , different re - orientational correlations times emerge for the four groups of nonequivalent molecules in the system : two times are enough to characterize three of the four groups and one time is associated to the remaining group .
the complete understanding of glassy systems , even in the case of the simpler rotational glasses , remains elusive in the sense that no complete theoretical picture can be drawn yet .
nevertheless , there is a general agreement on the existence of two main universal features referred to @xmath3 and @xmath4 relaxation .
rotational glasses of the type cbr@xmath1cl@xmath2 , with n=1 , 2 , 3 are glass formers displaying fast and slow rotational modes as they approach to the glass transition temperature . in this paper
we show that ccl@xmath0 has essentially the same dynamical behavior , as a function of temperature , to that of its isostructural glass formers .
long molecular dynamics simulations trajectories , based on a simple pair - wise additive force - field , yield to results that are fully in line with those obtained with nqr experiments .
the simulations clearly show that there are preferential axes of rotation , which are fixed with respect to the crystal orientation .
two of the inequivalent group of molecules are significantly faster than the other two , leading to a clear heterogeneity in the dynamics of the system .
moreover , it is found that the orientation of the two fast axes of rotations is the same , suggesting an overall dynamics anisotropy correlated to the molecular orientations .
finally , our results on ccl@xmath0 suggest that the observed heterogeneous dynamic of molecules of type cxcl@xmath120 with the same crystal structure could be due to the molecular environment on the crystal and not to the breaking of the tetrahedral symmetry of the molecule . then
, it would be interesting to corroborate if there is any relation between the fastest axis in the ccl@xmath0 molecules and the c - br axis in the cbrcl@xmath120 crystals .
we are performing md simulations to elucidate this open question .
, m.z . and
p. s acknowledge financial support of secytunc and conicet .
this work used computational resources from ccad universidad nacional de crdoba ( http://ccad.unc.edu.ar/ ) , in particular the mendieta cluster , which is part of sncad mincyt , repblica argentina .
a more detailed explanation of the method for counting molecular jumps and c - c radial distribution functions at the lowest and highest simulated temperatures along with experimental c - c distances are presented as supplemental material .
99 schneider , u. ; brand , r. ; lunkenheimer , p. ; loidl , a. excess wing in the dielectric loss of glass formers : a johari - goldstein @xmath4-relaxation ? .
_ , * 2000 * , 84 , 5560 - 5563 .
lunkenheimer , p. ; loidl , a. dielectric spectroscopy of glass - forming materials : @xmath3-relaxation and excess wing .
_ , * 2002 * , 284 , 205 - 219 .
affouard , f. ; cochin , e. ; dande , f ; decressain , r. ; descamps , m. ; haeussler , w. onset of slow dynamics in difluorotetrachloroethane glassy crystal .
_ j. chem .
_ , * 2005 * , 123 , 084501 .
ngai , k. l. ; paluch , m. classification of secondary relaxation in glass - formers based on dynamic properties .
_ j. chem .
_ , * 2004 * , 120 , 857 - 873 . pardo , l. c. ; lunkenheimer , p. ; loidl , a. @xmath3 and @xmath4 relaxation dynamics of a fragile plastic crystal .
_ j. chem .
_ , * 2006 * , 124 , 124911 .
johari , g. p. ; goldstein m. viscous liquids and the glass transition .
secondary relaxations in glasses of rigid molecules _ j. chem
_ , * 1970 * , 53 , 2372 .
ngai , k. l. correlation between the secondary @xmath4-relaxation time at @xmath123 with the kohlrausch exponent of the primary @xmath3 relaxation or the fragility of glass - forming materials .
e _ , * 1998 * , 57 , 7346 - 7349 .
schneider , u. ; brand , r. ; lunkenheimer , p. ; loidl , a. excess wing in the dielectric loss of glass formers : a johari - goldstein @xmath124 relaxation ? .
_ , * 2000 * , 84 , 5560 - 5563 .
johari , g. p. ; power , g. ; vij , j. k. localized relaxation strength and its mimicry of glass - softening thermodynamics .
_ j. chem .
_ , * 2002 * , 116 , 5908 - 5909 .
johari , g. p. intrinsic mobility of molecular glasses .
_ j. chem .
_ , * 1973 * , 58 , 1766 - 1770 .
capaccioli , s. ; ngai , k. l. ; shinyashiki , n. the johari - goldstein @xmath4-relaxation of water .
_ j. chem .
b _ , * 2007 * , 111 , 8197 - 8209 .
capaccioli , s. ; ngai , k. l. resolving the controversy on the glass transition temperature of water ? .
_ j. chem .
_ , * 2011 * , 135 , 104504 .
jimnez - ruiz , m. ; criado , a. ; bermejo , f.j . ;
cuello , g.j . ; trouw , f.r . ; fernndez - perea , r. ; lwen , h. ; cabrillo , c. ; fischer , h. e. purely dynamical signature of the orientational glass transition .
_ , * 1999 * , 83 , 2757 - 2760 .
jimnez - ruiz , m. ; gonzlez , m. a. ; bermejo , f.j . ; miller , m. a. ; birge , n. o. ; cendoya , i. ; alegra , a. relaxational dynamics in the glassy , supercooled liquid , and orientationally disordered crystal phases of a polymorphic molecular material .
b _ , * 1999 * , 59 , 9155 - 9166 .
voguel , m. ; rossler , e. on the nature of slow @xmath4-process in simple glass formers : a @xmath125h nmr study .
_ j. phys .
chem . b. _
, * 2000 * , 104 , 4285 - 4287 .
voguel , m . ; rossler , e. slow @xmath4 process in simple organic glass formers studied by one- and two - dimensional 2h nuclear magnetic resonance .
i. _ j. chem .
_ , * 2001 * , 114 , 5802 .
romanini , m. ; negrier , ph . ; tamarit , j. ll .
; capaccioli , s. ; barrio , m. ; pardo , l. c. ; mondieig , d. emergence of glassy - like dynamics in an orientationally ordered phase .
b _ , * 2012 * , 85 , 134201 .
drozd - rzoska , a. ; rzoska , s.j . ;
pawlus , s. ; tamarit , j. l. dynamics crossover and dynamic scaling description in vitrification of orientationally disordered crystal .
b _ , * 2006 * , 73 , 224205 .
martinez - garcia , j.c .
; tamarit , j. l. ; rzoska , s. j. prevalence for the universal distribution of relaxation times near the glass transitions in experimental model systems : rodlike liquid crystals and orientationally disordered crystals .
_ j. chem .
_ , * 2011 * , 134 , 144505 .
martinez - garcia , j.c .
; tamarit , j. l. ; rzoska , s. j. enthalpy space analysis of the evolution of the primary relaxation time in ultraslowing systems .
_ j. chem .
_ , * 2011 * , 134 , 024512 .
martinez - garcia , j.c .
; tamarit , j. l. ; capaccioli , s. ; barrio , m. ; veglio , n. ; pardo , l. c. -relaxation dynamics of orientanionally disordered mixed crystals composed of cl - adamantane and cn - adamantane .
_ j. chem .
_ , * 2010 * , 132 , 164516 .
caballero , n. b. ; zuriaga , m. ; carignano , m. ; serra , p. the plastic and liquid phases of ccl@xmath120br studied by molecular dynamics simulations .
_ j. chem .
_ , * 2012 * , 136 , 094515 .
zuriaga , m. j. ; perez , s. c. ; pardo , l. c. ; tamarit , j. ll .
dynamic heterogeneity in the glass - like monoclinic phases of cbr@xmath126cl@xmath2 , n = 0,1,2 .
_ j. chem .
_ , * 2012 * , 137 , 054506 .
zuriaga , m. ; pardo , l.c . ; lunkenheimer , p. ; tamarit , j. ll . ; veglio , n. ; barrio , m. ; bermejo , f.j .
; loidl , a. new microscopic mechanism for secondary relaxation in glasses .
_ , * 2009 * , 103 , 075701 .
pardo , l. c. ; barrio , m. ; tamarit , j. l. ; lpez , d. o. ; salud , j. ; negrier , p. ; mondieig , d. miscibility study in stable and metastable orientational disordered phases in a two - component system ( ch@xmath120)ccl@xmath120+ccl@xmath0 .
_ , * 1999 * , 308 , 204 - 210 .
pardo , l. c. ; barrio , m. ; tamarit , j. l. ; lpez , d. o. ; salud , j. ; negrier , p. ; mondieig , d. first experimental demonstration of crossed isodimorphism : ( ch@xmath120)@xmath120ccl + ccl@xmath0 melting phase diagram . _
_ , * 2001 * , 3 , 2644 - 2649 .
pardo , l. c. ; barrio , m. ; tamarit , j. l. ; negrier , p. ; lpez , d. o. ; salud , j. ; mondieig , d. stable and metastable phase diagram of the two - component system ( ch@xmath120)@xmath120ccl-(ch@xmath120)ccl@xmath120 : an example of crossed isodimorphism .
b _ , * 2001 * , 105 , 10326 - 10334 .
pardo , l. c. ; barrio , m. ; tamarit , j. l. ; lpez , d. o. ; salud , j. ; negrier , p. ; mondieig , d. stable and metastable orientationally disordered mixed crystals of the two - component system ( ch@xmath120)@xmath119ccl@xmath119+ccl@xmath0 . _
_ , * 2000 * , 321 , 438 - 444 .
pothoczki , sz . ; ottochian , a. ; rovira - esteva , m. ; pardo , l. c. ; tamarit , j. ll . ; cuello , g. j. role of steric and electrostatic effects in the short - range order of quasitetrahedral molecular liquids .
b. _ , * 2012 * , 85 , 014202 .
ohta , t. ; yamamuro , o. ; matsuo , t. heat capacities and phase transitions of cbrcl@xmath120 and cbr@xmath119cl@xmath119 .
_ j. phys .
_ , * 1995 * , 99 , 2403 - 2407 .
binbrek , o. s. ; lee - dadswell , s. e. ; torrie , b. h. ; powell , b. m. crystal structures of dibromodichloromethane and bromotrichloromethane .
_ , * 1999 * , 96 , 785 - 794 .
zuriaga , m. ; carignano , m. ; serra , p. rotational relaxation characteristics of the monoclinic phase of ccl@xmath0 .
_ j. chem .
_ , * 2011 * , 135 , 044504 .
cohen , s. ; powers , r. ; rudman , r. polymorphism of the crystalline methylchloromethane compounds .
vi . the crystal and molecular structure of ordered carbon tetrachloride . _ acta cryst .
b _ , * 1979 * , 35 , 1670 - 1674 .
carter , n. j. ; cross a. mechanics of the kinesin step .
_ nature _ , * 2005 * , 435 , 308 .
alexander , s. ; tzalmona , a. measurement of molecular rotation by @xmath127 nuclear quadrupole resonance relaxation times .
_ , * 1964 * , 13 , 546 - 547 .
in order to detect the rotational jumps of the ccl@xmath0 molecules we perform a detailed analysis of the trajectories based on an algorithm published by n. j. carter and a. cross ; _ nature _ * 435 * 308 ( 2005 ) .
let us consider the coordinates of each one of the cl atoms , relative to their bonded c atom , and denote them by @xmath128 . here
@xmath3 represent any of the cartesian coordinates @xmath20 , @xmath22 or @xmath21 ; the index @xmath81 denote molecule and @xmath129 refers to each one of the cl atoms of molecule @xmath81 .
@xmath130 where the angular brackets represent time average , the subindices @xmath131 and @xmath132 indicate the lapse of length @xmath133 before and after time @xmath134 , respectively ; and @xmath135 and @xmath136 are the corresponding standard deviations .
whenever there is a rotational jump the mean values and standard deviation are affected .
consequently , the test function @xmath137 reflects those changes with a spike
. the spike could be upwards or downwards depending on the direction of change of the average coordinate .
+ in order to have a single function per molecule , the twelve individual @xmath138 for an individual molecule can be combined in the following way : @xmath139 by careful inspection of the simulation trajectories and the response of the test functions we selected a time lapse @xmath133= 50 ps . in figure [ test ]
we show an example of the variation of the relative coordinates @xmath140 corresponding to the four cl atoms of molecule @xmath81 , which undergoes a jump at @xmath1412971.39 ns .
the figure also shows the corresponding four individual response functions of each cl atom : @xmath142 along with the final curve of the test function @xmath143 .
a threshold value of @xmath143=50 was used to determine the rotational jumps .
top : variation of the relative x coordinates of the four chlorine atoms of molecule @xmath81 in green ( cl@xmath116 ) , cyan ( cl@xmath119 ) , light green ( cl@xmath120 ) and orange ( cl@xmath0 ) .
the four individual response functions are also shown in the corresponding color , along with the final curve of the test function ( in black ) .
the threshold value of 50 is shown in red.,scaledwidth=65.0% ]
in order to analyze the crystalline structure of the simulation supercell , we calculated the carbon - carbon radial distribution function for each possible combination of carbon atoms within the different 4 groups of molecules in the system , at the all the simulation temperatures .
we observed a good agreement of the function s peaks with the experimental values for the carbon - carbon distances on the monoclinic cell . on figure [ rdf ]
we show the results at the lowest and highest simulation temperatures . | carbon tetrachloride ( ccl@xmath0 ) is one of the simplest compounds having a translationally stable monoclinic phase while exhibiting a rich rotational dynamics below 226 k. recent nuclear quadrupolar resonance ( nqr ) experiments revealed that the dynamics of ccl@xmath0 is similar to that of the other members of the isostructural series cbr@xmath1cl@xmath2 , suggesting that the universal relaxation features of canonical glasses such as @xmath3- and @xmath4-relaxation are also present in non - glass formers . using molecular dynamics ( md ) simulations we studied the rotational dynamics in the monoclinic phase of ccl@xmath0 .
the molecules undergo @xmath5 type jump - like rotations around each one of the four c - cl bonds .
the rotational dynamics is very well described with a master equation using as the only input the rotational rates measured from the simulated trajectories .
it is found that the heterogeneous dynamics emerges from faster and slower modes associated with different rotational axes , which have fixed orientations relative to the crystal and are distributed among the four non - equivalent molecules of the unit cell . |
the global and local analysis of solar acoustic modes has shown that the mode amplitudes and lifetimes are anti - correlated with the solar activity level , and depend strongly on the local magnetic flux @xcite .
most of the analyses have been done for average quantities without discrimination between quiet and active areas , while the solar cycle behavior of the quiet - sun regions is not well understood .
high - degree _ p_-mode lifetimes measured in the quiet - sun regions using cross - correlation analysis show variations with the activity cycle @xcite . in the quiet sun , no large - scale magnetic field concentrations , one of the potential damping sources , are present on the surface
. another explanation could be the activity - related variation of the convective properties near the solar surface .
the studies of the solar cycle variations of the size of the solar granules so far arrive at contradictory conclusions @xcite .
an attempt to mask strong surface activity and analyze high - degree _ p_-mode amplitudes in the quiet - sun regions at solar minimum and maximum indicated that the amplitude at solar minimum is higher than that at solar maximum @xcite , however , the effect introduced by the mask needs to be better understood . in this work
we apply ring - diagram analysis to eight years of data to characterize the high - degree acoustic mode amplitudes and widths in the quiet and active sun during declining phase of the activity cycle .
the mode parameters of the solar acoustic oscillations analyzed in this work are obtained from global oscillation network group ( gong ) and michelson doppler imager ( mdi ) high - resolution dopplergrams for the period from 2001 to 2009 using the standard ring - diagram technique @xcite of the gong and mdi pipelines .
the size of a standard patch in the ring analysis is @xmath0 .
the most distant patches are centered at [email protected] from the disk center .
more recently the pipeline developed for the helioseismic and magnetic imager ( hmi ) @xcite , applying both asymmetric @xcite and symmetric @xcite profiles fitting to derive the mode parameters , has also been used to analyze the mdi data , in the traditional mdi patch set .
the results from both procedures are consistent , thus , here we show the symmetric profile fitting results only .
the magnetic activity index ( mai ) values are computed from mdi 96-minute magnetograms for the same time frame of dopplergrams .
the mai is computed by averaging unsigned flux above 50 g and below 500 g to avoid contamination from noise in the measurements of the quiet - sun flux values and saturation issues of the mdi magnetograms @xcite .
we define the quiet - sun regions as those with mai values below 5 g. as discussed in @xcite , the amplitudes and widths fitted from the ring analysis need to be corrected for center - to - limb , residual b@xmath2-angle variations , and duty - cycle dependences . in our analysis
, gong data were restricted to duty cycle of 70@xmath3 and higher .
in contrast , mdi data have high duty cycle and therefore no selection criteria were used to reject the fitted parameteres .
in this work , only disk center patches were analyzed to avoid the foreshortening effects .
the results for amplitudes and widths shown in this paper are for a multiplet @xmath4 = 440 , @xmath5 = 2 ( @xmath6 = 3.2 mhz ) .
we find that the correlation of this multiplet with other multiplets in the [email protected] mhz frequency range over solar activity cycle is higher than 70@xmath3 .
the variations of amplitude and width as a function of time and mai , computed from all patches and the quiet - sun patches at the disk center using gong data , are shown in figure 1 . according to the mt.wilson sunspot index data ( see figure 2 ) , during most of the year 2008
the sun was quiet with no or very few small sunspots appearing on the solar surface .
therefore , we took average of year 2008 consisting of 13 carrington rotations as a reference , and plotted the mode parameters relative to this mean value .
the amplitudes increase by @xmath8 from the high solar activity period in 2001 till 2004 . after this period
, we find a long - term variation , but no clear association with activity cycle .
it is interesting to note that the mode amplitude , derived from variability of solar irradiance and gravity oscillations ( virgo ) and gong data using global helioseismology technique , shows similar wiggles though they are small in amplitude relative to the extent of the solar cycle trend @xcite .
if we ignore the unexplained long - term variations , we can conclude that the mode amplitudes obtained in our study increased from 2001 to 2008 by @xmath9 in all patches and by @xmath10 in the quiet - sun regions .
the widths decreased from 2001 to about 2008 by @xmath11 , then started showing a rising trend .
the variations in amplitude and width from mdi data ( see figure 1 , top panels ) show a solar cycle trend similar to that from gong data .
the long - term variations in the amplitudes obtained from gong data are not seen in the mdi amplitudes
. however , this could be due to the limited amount of data ( only 135 days ) that were analyzed with the mdi dopplergrams .
we plan to improve statistics on mdi data to confirm these results .
the decrease in amplitudes and increase in widths with increase in mai ( bottom panels of figure 1 ) are consistent with known results from global and local analysis .
the linear correlation analysis shows that the correlation coefficient between amplitude and mai is @xmath12 for all patches and @xmath13 for the quiet - sun patches .
the correlation coefficient between width and mai is 0.73 for all patches and 0.48 for the quiet - sun patches .
figure 2 shows the variation of the magnetic activity index over time .
the mai values of all patches at the disk center are well correlated with mt .
wilson sunspot index .
a weaker solar - cycle trend is also visible in the magnetic indices of the quiet - sun patches .
the solar cycle variation of the mais of the quiet - sun regions computed from mdi magnetograms was also noted in @xcite and could be due solar cycle variation of the strong - field component of the quiet - sun network @xcite .
this suggests that the magnetic field plays a role in the activity related variations of the acoustic mode parameters in the quiet sun . in an attempt to understand the long - term variations , on top of the solar cycle trend , seen in the amplitudes computed from gong data , we restrict our analysis to 2 g @xmath14 mai @xmath14 3 g patches .
we notice the long - term variation in the amplitude but no dependence on magnetic activity .
this leads to the conclusion that the variations are probably not related to solar activity cycle .
moreover , according to our preliminary estimates from the patches at higher latitudes , the variations are present up to the regions centered at [email protected] in latitude .
we plan to investigate this aspect in more detail after applying the necessary geometrical corrections .
this work utilizes data obtained by the global oscillation network group ( gong ) program , managed by the national solar observatory , which is operated by aura , inc . under a cooperative agreement with the national science foundation .
the data were acquired by instruments operated by the big bear solar observatory , high altitude observatory , learmonth solar observatory , udaipur solar observatory , instituto de astrofisica de canarias , and cerro tololo interamerican observatory .
soho is a mission of international cooperation between esa and nasa . | we study temporal variations in the amplitudes and widths of high - degree acoustic modes in the quiet and active sun by applying ring - diagram technique to the gong+ and mdi dopplergrams during the declining phase of cycle 23 .
the increase in amplitudes and decrease in line - widths in the declining phase of the solar activity is in agreement with previous studies .
a similar solar cycle trend in the mode parameters is also seen in the quiet - sun regions but with a reduced magnitude .
moreover , the amplitudes obtained from gong+ data show long - term variations on top of the solar cycle trend . |
resampling methods for dependent data such as time series have been studied extensively over the last decades . for an overview of existing bootstrap methods
see the monograph of @xcite and the review papers by @xcite , @xcite , @xcite , @xcite or the recent review paper by @xcite . among the most popular bootstrap procedures in time series analysis , we mention the autoregressive ( ar ) sieve bootstrap [ cf .
@xcite ( @xcite ) , @xcite , @xcite ] and block bootstrap and its variations ; cf .
@xcite , @xcite , @xcite ( @xcite ) , etc .
a recent addition to the available time series bootstrap methods was the linear process bootstrap ( lpb ) introduced by @xcite who showed its validity for the sample mean for univariate stationary processes without actually assuming linearity of the underlying process .
the main idea of the lpb is to consider the time series data of length @xmath0 as one large @xmath0-dimensional vector and to estimate appropriately the entire covariance structure of this vector .
this is executed by using tapered covariance matrix estimators based on flat - top kernels that were defined in @xcite .
the resulting covariance matrix is used to whiten the data by pre - multiplying the original ( centered ) data with its inverse cholesky matrix ; a modification of the eigenvalues , if necessary , ensures positive definiteness .
this decorrelation property is illustrated in figures 5 and 6 in @xcite .
after suitable centering and standardizing , the whitened vector is treated as having independent and identically distributed ( i.i.d . ) components with zero mean and unit variance .
finally , i.i.d . resampling from this vector and pre - multiplying the corresponding bootstrap vector of residuals with the cholesky matrix itself results in a bootstrap sample that has ( approximately ) the same covariance structure as the original time series . due to the use of flat - top kernels with compact support , an abruptly dying - out autocovariance structure
is induced to the bootstrap residuals .
therefore , the lpb is particularly suitable for but not limited to time series of moving average ( ma ) type . in a sense , the lpb could be considered the closest analog to an ma - sieve bootstrap which is not practically feasible due to nonlinearities in the estimation of the ma parameters .
a further similarity of the lpb to ma fitting , at least in the univariate case , is the equivalence of computing the cholesky decomposition of the covariance matrix to the innovations algorithm ; cf .
@xcite , @xcite and @xcite , the latter addressing the multivariate case .
typically , bootstrap methods extend easily from the univariate to the multivariate case , and the same is true for time series bootstrap procedures such as the aforementioned ar - sieve bootstrap and the block bootstrap . by contrast
, it has not been clear to date if / how the lpb could be successfully applied in the context of multivariate time series data ; a proposal to that effect was described in @xcite who refer to an earlier preprint of the paper at hand
but it has been unclear to date whether the multivariate lpb is asymptotically consistent and/or if it competes well with other methods .
here we attempt to fill this gap : we show how to implement the lpb in a multivariate context and prove its validity for the sample mean and for spectral density estimators , the latter being a new result even in the univariate case .
note that the limiting distributions of the sample mean and of kernel spectral density estimators depend only on the second - order moment structure . hence it is intuitive that the lpb would be well suited for such statistics since it generates a linear process in the bootstrap world that mimics well the second - order moment structure of the real world .
furthermore , in the spirit of the times , we consider the possibility that the time series dimension is increasing with sample size and identify conditions under which the multivariate linear process bootstrap ( mlpb ) maintains its asymptotic validity , even in this case .
the key here is to address the subject of consistently estimating the autocovariance sequence ; this is a sequence of matrices that we conveniently stack into one huge matrix .
we are then able to show consistency of an estimator based on the aforementioned flat - top tapers ; most importantly , the consistency holds true even when the time series dimension is allowed to increase with the sample size .
the paper is organized as follows . in section [ preliminaries ]
, we introduce the notation of this paper , discuss tapered covariance matrix estimation for multivariate stationary time series and state assumptions used throughout the paper ; we then present our results on convergence with respect to operator norm of tapered covariance matrix estimators .
the mlpb bootstrap algorithm and some remarks can be found in section [ bootstrapscheme ] , and results concerned with validity of the mlpb for the sample mean and kernel spectral density estimates are summarized in section [ asymptoticresults ] .
asymptotic results established for the case of increasing time series dimension are stated in section [ asymptoticresultsincreasing ] , where operator norm consistency of tapered covariance matrix estimates and a validity result for the sample mean are discussed .
a finite - sample simulation study is presented in section [ secsim ] . finally ,
all proofs , some additional simulations and a real data example on the weighted mean of an increasing number of stock prices taken from the german stock index dax can be found at the paper s supplementary material [ @xcite ] , which is also available at http://www.math.ucsd.edu/\textasciitilde politis / paper / mlpbsupplement.pdf[http://www.math.ucsd.edu/\textasciitilde politis / paper / mlpbsupplement.pdf ] .
suppose we consider an @xmath1-valued time series process @xmath2 with @xmath3 , and we have data @xmath4 at hand .
the process @xmath5 is assumed to be strictly stationary and its @xmath6 autocovariance matrix @xmath7 at lag @xmath8 is @xmath9 where @xmath10 , and the sample autocovariance @xmath11 at lag @xmath12 is defined by @xmath13 where @xmath14 is the @xmath15-variate sample mean vector .
here and throughout the paper , all matrix - valued quantities are written as bold letters , all vector - valued quantities are underlined , @xmath16 indicates the transpose of a matrix @xmath17 , @xmath18 the complex conjugate of @xmath17 and @xmath19 denotes the transposed conjugate of @xmath17 .
note that it is also possible to use unbiased sample autocovariances , that is , having @xmath20 instead of @xmath0 in the denominator of ( [ samplecovariance ] ) .
usually the biased version as defined in ( [ samplecovariance ] ) is preferred because it guarantees a positive semi - definite estimated autocovariance function , but our tapered covariance matrix estimator discussed in section [ taperedestimator ] is adjusted in order to become positive definite in any case .
now , let @xmath21 be the @xmath22-dimensional vectorized version of the @xmath23 data matrix @xmath24 $ ] , and denote the covariance matrix of @xmath25 , which is symmetric block toeplitz , by @xmath26 , that is , @xmath27 where @xmath28 is the covariance between the @xmath29th and @xmath30th entry of @xmath25 .
note that the second order stationarity of @xmath31 does _ not _ imply second - order stationary behavior of the vectorized @xmath22-dimensional data sequence @xmath25 .
this means that the covariances @xmath32 truly depend on both @xmath29 and @xmath30 and not only on the difference @xmath33 .
however , the following one - to - one correspondence between @xmath34 and @xmath35 holds true .
precisely , we have @xmath36 where @xmath37 and @xmath38 with @xmath39 and @xmath40 , and @xmath41 denotes the smallest integer greater or equal to @xmath42 .
if one is interested in estimating the quantity @xmath26 , it seems natural to plug in the sample covariances @xmath43 and @xmath44 in @xmath26 and to use @xmath45 but unfortunately this estimator is _ not _ a consistent estimator for @xmath26 in the sense that the operator norm of @xmath46 does not converge to zero .
this was shown by @xcite , and to dissolve this problem in the univariate case , they proposed a banded estimator of the sample covariance matrix to achieve consistency .
this has been generalized by @xcite , who considered general flat - top kernels as weight functions . in section [ taperedestimator ] ,
we follow the paper of @xcite and propose a tapered estimator of @xmath47 and show its consistency in theorem [ operatornormconvergence1 ] for the case of multivariate processes . moreover , we state a modified estimator that is guaranteed to be positive definite for any finite sample size and show its consistency in theorem [ operatornormconvergence2 ] and of related quantities in corollary [ operatornormconvergence3 ] . but prior to this , we state the assumptions that are used throughout this paper in the following .
@xmath31 is an @xmath1-valued strictly stationary time series process with mean @xmath48 and autocovariances @xmath49 defined in ( [ covariance ] ) such that @xmath50 for some @xmath51 to be further specified .
let @xmath52 for some matrix @xmath53 .
there exists a constant @xmath54 such that for all @xmath55 , all @xmath56 with @xmath12 and all @xmath57 , we have @xmath58 where @xmath59
. there exists an @xmath60 large enough such that for all @xmath61 the eigenvalues @xmath62 of the @xmath63 covariance matrix @xmath26 are bounded uniformly away from zero .
define the projection operator @xmath64 for @xmath65 , and suppose that for all @xmath66 , we have @xmath67 and @xmath68 , respectively , for some @xmath69 to be further specified . for the sample mean , a clt holds true .
that is , we have @xmath70 where @xmath71 denotes weak convergence , @xmath72 is a normal distribution with zero mean vector and covariance matrix @xmath73 with @xmath74 positive definite . for kernel
spectral density estimates @xmath75 as defined in ( [ fhat ] ) in section [ asymptoticresults ] , a clt holds true .
that is , for arbitrary frequencies @xmath76 , we have that @xmath77 converges to an @xmath78-dimensional normal distribution for @xmath79 and @xmath80 such that @xmath81 as @xmath82 , where the limiting covariance matrix is obtained from @xmath83 and the limiting bias from @xmath84 for all @xmath85 , where @xmath86 if @xmath87 and @xmath88 if @xmath89 and zero otherwise , respectively .
therefore , @xmath90 is assumed to be component - wise twice differentiable with lipschitz - continuous second derivatives .
assumption is quite standard , and the uniform convergence of sample autocovariances in ( a2 ) is satisfied under different types of conditions ( cf . [ remarka2 ] below ) and appears to be a crucial condition here .
the uniform boundedness of all eigenvalues away from zero in ( a3 ) is implied by a nonsingular spectral density matrix @xmath91 of @xmath92 .
this follows with ( [ gammadn ] ) and the inversion formula from @xmath93 for all @xmath94 , where @xmath95 and @xmath96 denotes the kronecker product .
the requirement of condition ( a3 ) fits into the theory for the univariate autoregressive sieve bootstrap as obtained in @xcite .
similarly , a nonsingular spectral density matrix @xmath91 implies positive definiteness of the long - run variance @xmath97 defined in ( a5 ) .
assumption is , for instance , fulfilled if the underlying process is linear or @xmath98-mixing with summable mixing coefficients by ibragimov s inequality ; cf .
, for example , @xcite , theorem 14.2 . to achieve validity of the mlpb for the sample mean and for kernel spectral density estimates in section [ asymptoticresults ] , we have to assume unconditional clts in ( a5 ) and ( a6 ) , which are satisfied also under certain mixing conditions [ cf .
@xcite , @xcite ] , linearity [ cf .
@xcite , @xcite ] or weak dependence [ cf .
note also that the condition @xmath81 includes the optimal bandwidth choice @xmath99 , @xmath100 for second - order kernels , which leads to a nonvanishing bias in the limiting normal distribution .
[ remarka2 ] assumption ( a2 ) is implied by different types of conditions imposed on the underlying process @xmath31 .
we present sufficient conditions for ( a2 ) under linearity , mixing or weak dependence type conditions .
more precisely , ( a2 ) is satisfied if the process @xmath31 fulfills one of the following conditions : _ linearity_. suppose the process is linear ; that is , @xmath101 , @xmath102 , where @xmath103 is an i.i.d .
white noise with finite fourth moments @xmath104 for all @xmath105 and the sequence of @xmath6 coefficient matrices @xmath106 is component - wise absolutely summable .
_ mixing - type condition_. let @xmath107 denote the @xmath108th order joint cumulant of @xmath109 [ cf.@xcite ] , and suppose @xmath110 for all @xmath57 .
note that this is satisfied if @xmath111 is @xmath98-mixing such that @xmath112 and @xmath113 for some @xmath114 ; cf .
@xcite , page 221 .
_ weak dependence - type condition_. suppose for all @xmath57 , we have @xmath115 where @xmath116 and @xmath117 is absolutely summable , that is , @xmath118 cf . @xcite . to adopt the technique of @xcite , let @xmath119 be a so - called _ flat - top taper _ [ cf .
@xcite ] , where @xmath120 and @xmath121 . the @xmath122-scaled version of @xmath123 is defined by @xmath124 for some @xmath125 .
as @xcite argues , it is advantageous to have a smooth taper @xmath126 , so the truncated kernel that corresponds to @xmath127 for all @xmath128 is not recommended .
the simplest example of a continuous taper function @xmath126 with @xmath129 is the trapezoid @xmath130 which is used in section [ secsim ] for the simulation study ; the trapezoidal taper was first proposed by @xcite in a spectral estimation setup .
observe also that the banding parameter does not need to be an integer .
the tapered estimator @xmath131 of @xmath26 is given by @xmath132 where @xmath133 and @xmath134 .
the following theorem [ operatornormconvergence1 ] deals with consistency of the tapered estimator @xmath135 with respect to operator norm convergence .
it extends theorem 1 in @xcite to the multivariate case and does not rely on the concept of physical dependence only .
the operator norm of a complex - valued @xmath6 matrix @xmath17 is defined by @xmath136 and it is well known that @xmath137 , where @xmath138 denotes the largest eigenvalue of a matrix @xmath139 ; cf .
@xcite , page 296 .
[ operatornormconvergence1 ] suppose that assumptions with @xmath140 and are satisfied .
then it holds @xmath141\\[-8pt]\nonumber & & \qquad \leq \frac{4md^2(\lfloor c_\kappa l\rfloor+1)}{\sqrt{n}}+2\sum _ { h=0}^{\lfloor c_\kappa l\rfloor}\frac{{\vert}h{\vert}}{n}\bigl{\vert}\mathbf{c}(h)\bigr{\vert}_1 + 2\sum_{h = l+1}^{n-1}\bigl{\vert}\mathbf{c}(h)\bigr{\vert}_1.\end{aligned}\ ] ] the second term on the right - hand side of ( [ convergenceeq ] ) can be represented as @xmath142 and vanishes asymptotically due to the kronecker lemma and is of order @xmath143 .
the third one converges to zero for @xmath144 as @xmath82 and the leading first term for @xmath145 . hence , for @xmath146 , the right - hand side of ( [ convergenceeq ] ) vanishes asymptotically .
however , if @xmath147 for @xmath148 for some @xmath149 , setting @xmath150 fixed suffices . in this case
, the expression on the right - hand side of ( [ convergenceeq ] ) is of faster order @xmath151 . as already pointed out by @xcite , the tapered estimator @xmath152 is not guaranteed to be positive semi - definite or even to be positive definite for finite sample sizes .
however , @xmath152 is at least `` asymptotically positive definite '' under assumption ( a3 ) and due to ( [ convergenceeq ] ) if @xmath146 holds . in the following , we require a consistent estimator for @xmath26 which is positive definite for all finite sample sizes to be able to compute its cholesky decomposition for the linear process bootstrap scheme that will be introduced in section [ bootstrapscheme ] below . to obtain an estimator of @xmath26 related to @xmath131 that is assured to be positive definite for all sample sizes , we construct a modified estimator @xmath153 in the following .
let @xmath154 be the diagonal matrix of sample variances , and define @xmath155 .
now we consider the spectral factorization @xmath156 , where @xmath157 is an @xmath158 orthogonal matrix and @xmath159 is the diagonal matrix containing the eigenvalues of @xmath160 such that @xmath161 .
it is worth noting that this factorization always exists due to symmetry of @xmath162 , but that the eigenvalues can be positive , zero or even negative .
now , define @xmath163 where @xmath164 and @xmath165 . here , @xmath166 and @xmath167 are user defined constants that ensure the positive definiteness of @xmath168 .
contrary to the univariate case discussed in @xcite , we propose to adjust the eigenvalues of the ( equivariant ) correlation matrix @xmath162 instead of @xmath152 , which then comes along without a scaling factor in the definition of @xmath169 .
further , note that setting @xmath170 leads to a positive semi - definite estimate if @xmath152 is indefinite , which does not suffice for computing the cholesky decomposition , and also that @xmath168 generally loses the banded shape of @xmath152 .
theorem [ operatornormconvergence2 ] below , which extends theorem 3 in @xcite , shows that the modification of the eigenvalues does affect the convergence results obtained in theorem [ operatornormconvergence1 ] just slightly .
[ operatornormconvergence2 ] under the assumptions of theorem [ operatornormconvergence1 ] , it holds @xmath171 in comparison to the upper bound established in theorem [ operatornormconvergence1 ] , two more terms appear on the right - hand side of ( [ convergenceeq2 ] ) which do converge as well to zero as @xmath0 tends to infinity .
note that the first three summands that the right - hand sides of ( [ convergenceeq ] ) and ( [ convergenceeq2 ] ) have in common , remain the leading terms if @xmath172 .
we also need convergence and boundedness in operator norm of quantities related to @xmath173 .
the required results are summarized in the following corollary .
[ operatornormconvergence3 ] under assumptions with @xmath140 , and , we have : @xmath174 and @xmath175 are terms of order @xmath176 , where @xmath177 and @xmath178 if @xmath146 .
@xmath179 and @xmath180 are of order @xmath181 and @xmath182 if @xmath183 , and holds for some @xmath184 .
@xmath185 , @xmath186 , @xmath187 , @xmath188 are bounded from above and below .
@xmath189 , @xmath190 and @xmath191 , @xmath192 are bounded from above and below ( in probability ) if @xmath178 and @xmath182 , respectively .
[ individualbanding ] in section [ taperedestimator ] , we propose to use a global banding parameter @xmath122 that down - weights the autocovariance matrices for increasing lag ; that is , the entire matrix @xmath49 is multiplied with the same @xmath193 in ( [ gammahat ] ) .
however , it is possible to use individual banding parameters @xmath194 for each sequence of entries @xmath195 , @xmath196 as proposed in @xcite , compare also the simulation section . to get a tapered estimate @xmath152 of the covariance matrix @xmath26
, some parameters have to be chosen by the practitioner .
these are the flat - top taper @xmath126 and the banding parameter @xmath122 , which are both responsible for the down - weighting of the empirical autocovariances @xmath197 with increasing lag @xmath56 .
to select a suitable taper @xmath126 from the class of functions ( [ kappa ] ) , we have to select @xmath121 and the function @xmath198 which determine the range of the decay of @xmath126 to zero for @xmath199 and its form over this range , respectively . for some examples of flat - top tapers ,
compare @xcite ( @xcite ) .
however , the selection of the banding parameter @xmath122 appears to be more crucial than choosing the tapering function @xmath126 among the family of well - behaved flat - top kernels as discussed in @xcite .
this is comparable to nonparametric kernel estimation where usually the bandwidth plays a more important role than the shape of the kernel .
we focus on providing an empirical rule for banding parameter selection that has already been used in @xcite for the univariate lpb .
they make use of an approach primarily proposed in @xcite to estimate the bandwidth in spectral density estimation which has been generalized to the multivariate case in @xcite . in the following , we adopt this technique based on the correlogram / cross - correlogram [ cf .
@xcite ( @xcite , section 6 ) ] for our purposes .
let @xmath200 be the sample ( cross-)correlation between the two univariate time series @xmath201 and @xmath202 at lag @xmath8 .
now , define @xmath203 as the smallest nonnegative integer such that @xmath204 for @xmath205 , where @xmath206 is a fixed constant , and @xmath207 is a positive , nondecreasing integer - valued function of @xmath0 such that @xmath208 .
note that the constant @xmath209 and the form of @xmath207 are the practitioner s choice . as a rule of thumb
, we refer to @xcite ( @xcite ) who makes the concrete recommendation @xmath210 and @xmath211 . after having computed @xmath212 for all @xmath213 ,
we take @xmath214 as a data - driven _ global _ choice of the banding parameter @xmath122 . by setting @xmath215
, we get data - driven _ individual _ banding parameter choices as discussed in remark [ individualbanding ] . for theoretical justification of this empirical selection of a global cut - off point as the maximum over individual choices and assumptions that lead to successful adaptation
, we refer to theorem 6.1 in @xcite .
note also that for positive definite covariance matrix estimation , that is , for computing @xmath168 , one has to select two more parameters @xmath216 and @xmath217 , which have to be nonnegative and might be set equal to one as suggested in @xcite .
in this section , we describe the multivariate linear process bootstrap ( mlpb ) in detail , discuss some modifications and comment on the special case where the tapered covariance estimator becomes diagonal .
let @xmath218 be the @xmath23 data matrix consisting of @xmath1-valued time series data @xmath219 of sample size @xmath0 .
compute the centered observations @xmath220 , where @xmath221 , let @xmath222 be the corresponding @xmath23 matrix of centered observations , and define @xmath223 to be the @xmath22-dimensional vectorized version of @xmath222 .
compute @xmath224 , where @xmath225 denotes the lower left triangular matrix @xmath226 of the cholesky decomposition @xmath227 .
let @xmath228 be the standardized version of @xmath229 , that is , @xmath230 , @xmath231 , where @xmath232 and @xmath233 .
generate @xmath234 by i.i.d .
resampling from @xmath235 .
compute @xmath236 , and let @xmath237 be the matrix that is obtained from @xmath238 by putting this vector column - wise into an @xmath23 matrix , and denote its columns by @xmath239 . regarding steps 3 and 4 above and due to the multivariate nature of the data
, it appears to be even more natural to split the @xmath22-dimensional vector @xmath228 in step 3 above in @xmath0 sub - vectors , to center and standardize them and to apply i.i.d .
resampling to these vectors to get @xmath240 .
more precisely , steps 3 and 4 can be replaced by : let @xmath241 be the standardized version of @xmath229 , that is , @xmath242 , where @xmath243 , @xmath244 and @xmath245 .
generate @xmath246 by i.i.d .
resampling from @xmath247 .
this might preserve more higher order features of the data that are not captured by @xmath168 .
however , comparative simulations ( not reported in the paper ) indicate that the finite sample performance is only slightly affected by this sub - vector resampling .
[ remark2 ] if @xmath248 , the banded covariance matrix estimator @xmath152 ( and @xmath249 as well ) becomes diagonal . in this case and if steps 3@xmath250 and 4@xmath250 are used , the lpb as described above is equivalent to the classical i.i.d .
bootstrap . here , note the similarity to the autoregressive sieve bootstrap which boils down to an i.i.d . bootstrap if the autoregressive order is @xmath251 .
in this section , we establish validity of the mlpb for the sample mean .
the following theorem generalizes theorem 5 of @xcite to the multivariate case under somewhat more general conditions . [ validitysamplemean ] under assumptions for some @xmath184 , ,
, for @xmath252 , and @xmath183 , the mlpb is asymptotically valid for the sample mean @xmath253 , that is , @xmath254 and @xmath255 , where @xmath256 . the short - hand @xmath257 for @xmath258 is used to denote @xmath259 for all @xmath66 .
here we prove consistency of the mlpb for kernel spectral density matrix estimators ; this result is novel even in the univariate case .
let @xmath260 the periodogram matrix , where @xmath261 is the discrete fourier transform ( dft ) of @xmath262 , @xmath263 .
we define the estimator @xmath264 for the spectral density matrix @xmath265 , where @xmath266 is the integer part of @xmath42 , @xmath267 are the fourier frequencies , @xmath268 is the bandwidth and @xmath269 is a symmetric and square integrable kernel function @xmath270 that satisfies @xmath271 and @xmath272 and we set @xmath273 .
let @xmath274 be the bootstrap analogue of @xmath275 based on @xmath276 generated from the mlpb scheme and let @xmath277 be the bootstrap analogue of @xmath278 .
[ validitykernelspectral ] suppose assumptions with @xmath279 specified below , , , for @xmath280 and are satisfied . if @xmath79 and @xmath80 such that @xmath81 as well as @xmath281 and @xmath282 for some sequence @xmath283 , the mlpb is asymptotically valid for kernel spectral density estimates @xmath284 .
that is , for all @xmath285 and arbitrary frequencies @xmath286 ( not necessarily fourier frequencies ) , it holds @xmath287 where @xmath288 and , in particular , @xmath289 and @xmath290 , for all @xmath85 and all @xmath291 $ ] , respectively . for statistics @xmath292 contained in the broad class of functions of generalized means ,
@xcite discussed how by using a preliminary blocking scheme tailor - made for a specific statistic of interest , the mlpb can be shown to be consistent .
this class of statistics contains estimates @xmath292 of @xmath293 with @xmath294 such that @xmath295 for some sufficiently smooth functions @xmath296 , @xmath297 and fixed .
they propose to block the data first according to the known function @xmath198 and to apply then the ( m)lpb to the blocked data .
more precisely , the multivariate lpb - of - blocks bootstrap is as follows : define @xmath298 , and let @xmath299 be the set of blocked data .
apply the mlpb scheme of section [ bootstrapscheme ] to the @xmath108-dimensional blocked data @xmath300 to get bootstrap observations @xmath301 .
compute @xmath302 .
repeat steps 2 and 3 @xmath303-times , where @xmath303 is large , and approximate the unknown distribution of @xmath304 by the empirical distribution of @xmath305 .
the validity of the multivariate lpb - of - blocks bootstrap for some statistic @xmath292 can be verified by checking the assumptions of theorem [ validitysamplemean ] for the sample mean of the new process @xmath306 .
in this section , we consider the case when the time series dimension @xmath15 is allowed to increase with the sample size @xmath0 , that is , @xmath307 as @xmath82 . in particular , we show consistency of tapered covariance matrix estimates and derive rates that allow for an asymptotic validity result of the mlpb for the sample mean in this case .
the recent paper by @xcite gives a thorough discussion of the estimation of toeplitz covariance matrices for univariate time series . in their setup , that covers also the possibility of having multiple datasets from the same data generating process , @xcite establish the optimal rates of convergence using the two simple flat - top kernels discussed in section [ taperedestimator ] , namely the truncated ( i.e. , case of pure banding
no tapering ) and the trapezoid taper . when the strength of dependence is quantified via a smoothness condition on the spectral density ,
they show that the trapezoid is superior to the truncated taper , thus confirming the intuitive recommendations of @xcite .
the asymptotic theory of @xcite allows for increasing number of time series and increasing sample size , but their framework does not contain the multivariate time series case , neither for fixed nor for increasing time series dimension , which will be discussed in this section .
note that theorem 1 in @xcite for the univariate case , as well as our theorem [ operatornormconvergence1 ] for the multivariate case of fixed time series dimension , give upper bounds that are quite sharp , coming within a log - term to the ( gaussian ) optimal rate found in theorem 2 of @xcite . instead of assumptions ( a1)(a5 ) that have been introduced in section [ assumptions ] and used in theorem [ validitysamplemean ] to obtain bootstrap consistency for the sample mean for fixed dimension @xmath15
, we impose the following conditions on the sequence of time series process @xmath308 of now increasing dimension .
@xmath309 is a sequence of @xmath310-valued strictly stationary time series processes with mean vectors @xmath311 and autocovariances @xmath312 defined as in ( [ covariance ] ) . here
, @xmath313 is a nondecreasing sequence of positive integers such that @xmath314 as @xmath82 and , further , suppose @xmath315 for some @xmath279 to be further specified .
there exists a constant @xmath316 such that for all @xmath55 and all @xmath56 with @xmath12 , we have @xmath317 there exists an @xmath60 large enough such that for all @xmath61 and all @xmath318 the eigenvalues @xmath62 of the @xmath63 covariance matrix @xmath26 are bounded uniformly away from zero and from above .
define the sequence of projection operators @xmath319 for @xmath320 , and suppose @xmath321 and @xmath322 for the sample mean , a cramr wold - type clt holds true .
that is , for any real - valued sequence @xmath323 of @xmath324-dimensional vectors with @xmath325 for all @xmath55 and @xmath326 , we have @xmath327 assumptions ( a1@xmath250)(a4@xmath250 ) are uniform analogues of ( a1)(a4 ) , which are required here to tackle the increasing time series dimension @xmath15 . in particular , ( a1@xmath250 ) implies @xmath328 observe also that the autocovariances @xmath329 are assumed to decay with increasing lag @xmath56 , that is , in time direction , but they are not assumed to decay with increasing @xmath330 , that is , with respect to increasing time series dimension .
therefore , we have to make use of square summable sequences in ( a5@xmath250 ) to get a clt result .
this technique has been used , for example , by @xcite and @xcite to establish central limit results for the estimation of an increasing number of autoregressive coefficients .
a simple sufficient condition for ( a5@xmath250 ) is , for example , the case of @xmath331 being a sequence of i.i.d .
gaussian processes with eigenvalues of @xmath332 bounded uniformly from above and away from zero .
the following theorem generalizes the results of theorems [ operatornormconvergence1 ] and [ operatornormconvergence2 ] and of corollary [ operatornormconvergence3 ] to the case where @xmath333 is allowed to increase with the sample size .
in contrast to the case of a stationary spatial process on the plane @xmath334 ( where a data matrix is observed that grows in both directions asymptotically as in our setting ) , we do not assume that the autocovariance matrix decays in all directions .
therefore , to be able to establish a meaningful theory , we have to replace ( a1)(a5 ) by the uniform analogues ( a1@xmath335)(a5@xmath250 ) , and due to ( [ rate ] ) , an additional factor @xmath336 turns up in the convergence rate and has to be taken into account . [ operatornormconvergence3d ] under assumptions with @xmath279 specified below , and
, we have : @xmath174 and @xmath175 are terms of order @xmath337 , where @xmath338 and @xmath339 if @xmath340 .
@xmath179 and @xmath180 are both terms of order @xmath341 and @xmath342 if @xmath343 .
@xmath185 , @xmath186 , @xmath187 and @xmath344 are bounded from above and below . @xmath345 and
@xmath346 as well as @xmath347 and @xmath348 are bounded from above and below in probability if @xmath339 and @xmath349 , respectively .
the required rates for the banding parameter @xmath122 and the time series dimension @xmath15 to get operator norm consistency @xmath350 can be interpreted nicely .
if @xmath198 is chosen to be large enough , @xmath351 becomes the leading term , and there is a trade - off between capturing more dependence of the time series in time direction ( large @xmath122 ) and growing dimension of the time series in cross - sectional direction ( large @xmath15 ) . the subsequent theorem is a cramr wold - type generalization of theorem [ validitysamplemean ] to the case where @xmath333 is allowed to grow at an appropriate rate with the sample size . to tackle the increasing time series dimension and to prove such a clt result , we have to make use of appropriate sequences of square summable vectors @xmath352 as described in ( a5@xmath250 ) above .
[ validitysamplemeand ] under assumptions with @xmath279 specified below , , , for @xmath252 , as well as @xmath353 and @xmath354 for some sequence @xmath283 , the mlpb is asymptotically valid for the sample mean @xmath355 . that is , for any real - valued sequence @xmath323 of @xmath324-dimensional vectors with @xmath325 for all @xmath55 and @xmath356 , we have @xmath357 and @xmath358 . in practice
, the computational requirements can become very demanding for large @xmath15 and @xmath0 . in this case
, we suggest to split the data vector @xmath25 in few subsamples @xmath359 , say , and to apply the mlpb scheme to each subsample separately .
this operation can be justified by the fact that dependence structure is distorted only few times . precisely , we suggest the following procedure : for small @xmath360 , define @xmath361 and @xmath362 such that @xmath363 , and let @xmath364 , @xmath365 , where @xmath366 is filled up with zeros if @xmath367 .
apply the mlpb bootstrap scheme as described in section [ bootstrapscheme ] separately to the subsamples @xmath368 to get @xmath369 .
put @xmath370 end - to - end together , and discard the last @xmath371 values to get @xmath372 and @xmath373 . here ,
computationally demanding operations as eigenvalue decomposition , cholesky decomposition and matrix inversion have to be executed only for lower - dimensional matrices , such that the algorithm above is capable to reduce the computation time considerably .
further , to regain efficiency , we propose to use the pooled sample mean @xmath355 for centering and @xmath374 for whitening and re - introducing correlation structure for _ all _ subsamples in step 2 . here
, @xmath375 is obtained analogously to ( [ gammahatepsilon ] ) , but based on the upper - left @xmath376 sub - matrix of @xmath135 .
in this section we compare systematically the performance of the multivariate linear process bootstrap ( mlpb ) to that of the vector - autoregressive sieve bootstrap ( ar - sieve ) , the moving block bootstrap ( mbb ) and the tapered block bootstrap ( tbb ) by means of simulation . in order to make such a comparison ,
we have chosen a statistic for which all methods lead to asymptotically correct approximations .
being interested in the distribution of the sample mean , we compare the aforementioned bootstrap methods by plotting : root mean squared errors ( rmse ) for estimating the variances of @xmath377 and coverage rates ( cr ) of 95% bootstrap confidence intervals for the components of @xmath378 for two data generating processes ( dgps ) and three sample sizes in two different setups . first , in section [ secsimtuning ] , we compare the performance of all aforementioned bootstraps with respect to ( w.r.t . ) tuning parameter choice .
these are the banding parameter @xmath122 ( mlpb ) , the autoregressive order @xmath379 ( ar - sieve ) and the block length @xmath380 ( mbb , tbb ) .
furthermore , we report rmse and cr for data - adaptively chosen tuning parameters to investigate how accurate automatic selection procedures can work in practice .
second , in section [ secsimdimension ] , we investigate the effect of the time series dimension @xmath15 on the performance of the different bootstrap approaches . for each case , we have generated @xmath381 time series and @xmath382 bootstrap replications have been used in each step . for ( a ) , the exact covariance matrix of @xmath383 is estimated by 20,000 monte carlo replications .
further , we use the trapezoidal kernel defined in ( [ trapezoid ] ) to taper the sample covariance matrix for the mlpb and the blocks for the tbb . to correct the covariance matrix estimator @xmath135 to be positive definite , if necessary
, we set @xmath384 and @xmath385 to get @xmath386 .
this choice has already been used by @xcite and simulation results ( not reported in this paper ) indicate that the performance of the mlpb reacts only slightly to this choice .
we have used the sub - vector resampling scheme , that is , steps 3@xmath387 and 4@xmath387 described in section [ bootstrapscheme ] .
some additional simulation results and a real data application of the mlpb to the weighted mean of an increasing number of german stock prices taken from the dax index can be found in the supplementary material to this paper [ @xcite ] .
the r code is available at http://www.math.ucsd.edu/\textasciitilde politis / soft / function_mlpb.r[http://www.math.ucsd.edu/\textasciitilde politis / soft / function_mlpb.r ] .
we consider realizations @xmath388 of length @xmath389 from two bivariate ( @xmath390 ) dgps .
precisely , we study a first - order vector moving average process @xmath391 and a first - order vector autoregressive process @xmath392 where @xmath393 is a normally distributed i.i.d .
white noise process and @xmath394 have been used in all cases .
it is worth noting that ( asymptotically ) all bootstrap procedures under consideration yield valid approximations for both models above . for the vma(1 )
model , mlpb is valid for all ( sufficiently small ) choices of banding parameters @xmath395 , but ar - sieve is valid only asymptotically for @xmath396 tending to infinity at an appropriate rate with increasing sample size @xmath0 .
this relationship of mlpb and ar - sieve is reversed for the var(1 ) model . for the mbb and the tbb
, the block length has to increase with the sample size for both dgps . and cr of bootstrap confidence intervals for @xmath397 by mlpb ( solid ) , ar - sieve ( dashed ) , mbb ( dotted ) and tbb ( dash - dotted )
are reported vs. the respective tuning parameters @xmath398 for the vma(1 ) model with sample size @xmath399 .
line segments indicate results for data - adaptively chosen tuning parameters .
mlpb with individual ( grey ) and global ( black ) banding parameter choice are reported . ] , but with var(1 ) model . ]
in addition to the results for tuning parameters @xmath400 , we show also rmse and cr for tuning parameters chosen by automatic selection procedures in figures [ fig2 ] and [ fig3 ] . for the mlpb , we report results for data - adaptively chosen global and individual banding parameters as discussed in section [ selectionsec ] . for the ar - sieve , the order of the var model fitted to the data has been chosen by using the * r * routine var@xmath401 contained in the package * vars * with _ lag_.__max__@xmath402 .
the block length is chosen by using the * r * routine _ b_.__star__@xmath401 contained in the package * np*. in figures [ fig2 ] and [ fig3 ] , we report only the results corresponding to the first component of the sample mean , as those for the second component lead qualitatively to the same results . we show them in the supplementary material , which contains also corresponding simulation results for a normal white noise dgp . for data generated by the vma(1 ) model , figure [ fig2 ] shows that the mlpb outperforms ar - sieve , mbb and tbb for adequate tuning parameter choice , that is , @xmath403 . in this case
, the mlpb generally behaves superiorly , with respect to rmse and cr , to the other bootstrap methods for all tuning parameter choices of @xmath379 and @xmath380 .
this was not unexpected since , by design , the mlpb can approximate very efficiently the covariance structure of moving average processes .
nevertheless , due to the fact that all proposed bootstrap schemes are valid at least asymptotically , ar - sieve gets rid of its bias with increasing order @xmath379 , but at the expense of increasing variability and consequently also increasing rmse .
mlpb with data - adaptively chosen banding parameter performs quite well , where the individual choice tends to perform superiorly to the global choice in most cases . in comparison ,
mbb and tbb seem to perform quite well for adequate block length , but they lose in terms of rmse as well as cr performance if the block length is chosen automatically . the data from the var(1 ) model is highly persistent due to the coefficient @xmath404 near to unity .
this leads to autocovariances that are rather slowly decreasing with increasing lag and , consequently , to large variances of @xmath405 .
figure [ fig3 ] shows that ar - sieve outperforms mlpb , mbb and tbb with respect to cr for small ar orders @xmath406 .
this is to be expected since the underlying var(1 ) model is captured well by ar - sieve even with finite sample size .
but the picture appears to be different with respect to rmse .
here , mlpb may perform superiorly for adequate tuning parameter choice , but this effect can be explained by the very small variance that compensates its large bias , in comparison to the ar - sieve ( bias and variance not reported here ) leading to a smaller rmse .
this phenomenon is also illustrated by the poor performance of mlpb with respect to cr for small choices of @xmath122 .
however , more surprising is the rather good performance of the mlpb if the banding parameter is chosen data - adaptively , where the mlpb appears to be comparable to the ar - sieve in terms of rmse and is at least close with respect to cr .
further , as observed already for the vma(1 ) model in figure [ fig2 ] , the individual banding parameter choice generally tends to outperform the global choice here again .
similarly , it can be seen here that the performance of ar - sieve worsens with increasing @xmath379 at the expense of increasing variability .
the block bootstraps mbb and tbb appear to be clearly inferior to mlpb and ar - sieve , particularly with respect to cr , but also with respect to rmse if tuning parameters are chosen automatically .
we consider @xmath15-dimensional realizations @xmath407 with @xmath389 from two dgps of several dimensions .
precisely , we study first - order vector moving average processes @xmath408 and first - order vector autoregressive processes @xmath409 of dimension @xmath410 , where @xmath411 is a @xmath15-dimensional normally distributed i.i.d
. white noise process , and @xmath412 and @xmath413 are such that @xmath414 observe that the vma(1 ) and var(1 ) models considered in section [ secsimtuning ] are included in this setup for @xmath390 . in figures [ fig4 ] and [ fig5 ] , we compare the performance of mlpb , ar - sieve , mbb and tbb for the dgps above using rmse and cr averaged over all @xmath15 time series coordinates .
precisely , we compute rmse individually for the estimates of @xmath415 , @xmath66 and plot the averages in the upper half of figures [ fig4 ] and [ fig5 ] . similarly , we plot averages of individually calculated cr of bootstrap confidence intervals for @xmath416 , @xmath66 in the lower halfs .
all tuning parameters are chosen in a data - based and optimal way , as described in section [ secsimtuning ] , and to reduce computation time , the less demanding algorithm , as described in section [ secreduction ] with @xmath417 , is used . ,
@xmath66 and average cr of bootstrap confidence intervals for @xmath416 , @xmath66 , by mlpb ( solid ) , ar - sieve ( dashed ) , mbb ( dotted ) and tbb ( dash - dotted ) with data - based optimal tuning parameter choices are reported vs. the dimension @xmath418 for the vma@xmath419(1 ) model with sample size @xmath420 .
mlpb with individual ( grey ) and global ( black ) banding parameter choice are reported . ] , but with var@xmath419(1 ) model . ] for the vma(1 ) dgps in figure [ fig4 ] , the mlpb with individual banding parameter choice outperforms the other approaches essentially for all time series dimension under consideration with respect to averaged rmse and cr .
in particular , larger time series dimensions do not seem to have a large effect on the performance of all bootstraps for the vma(1 ) dgps , with the only exception being the mlpb with global banding parameter choice .
in particular , the latter is clearly inferior in comparison to the mlpb with individually chosen banding parameter , which might be explained by sparsity of the covariance matrix @xmath421 . in figure
[ fig5 ] , for the var(1 ) dgps , the picture is different from the vma(1 ) case above .
the influence of larger time series dimension on rmse ( and less pronounced for cr ) performance is much more pronounced and clearly visible .
in particular , the rmse blows up with increasing dimension @xmath15 for all four bootstrap methods , which is due to the also increasing variance of the process .
note that the zig - zag shape of the rmse curves is due to the back and forth switching from @xmath422 to @xmath423 on the diagonal of @xmath424 . as already observed for the vma(1 ) dgps , the mlpb with individual banding parameter choice again performs best over essentially all time series dimensions with respect to average rmse and average cr .
in particular , mlpb with individual choice is superior to the global choice . here
, the good performance of the mlpb is somewhat surprising as the var(1 ) dgps have rather slowly decreasing autocovariance structure , where we expected an ar - sieve to be more suitable .
the authors thank timothy mcmurry for his helpful advice on the univariate case and three anonymous referees and the editor who helped to significantly improve the presentation of the paper . | multivariate time series present many challenges , especially when they are high dimensional .
the paper s focus is twofold .
first , we address the subject of consistently estimating the autocovariance sequence ; this is a sequence of matrices that we conveniently stack into one huge matrix .
we are then able to show consistency of an estimator based on the so - called _ flat - top tapers _ ; most importantly , the consistency holds true even when the time series dimension is allowed to increase with the sample size .
second , we revisit the linear process bootstrap ( lpb ) procedure proposed by mcmurry and politis [ _ j .
time series anal . _
* 31 * ( 2010 ) 471482 ] for univariate time series . based on the aforementioned stacked autocovariance matrix estimator ,
we are able to define a version of the lpb that is valid for multivariate time series . under rather general assumptions ,
we show that our multivariate linear process bootstrap ( mlpb ) has asymptotic validity for the sample mean in two important cases : ( a ) when the time series dimension is fixed and ( b ) when it is allowed to increase with sample size . as an aside , in case ( a )
we show that the mlpb works also for spectral density estimators which is a novel result even in the univariate case .
we conclude with a simulation study that demonstrates the superiority of the mlpb in some important cases
. ./style / arxiv - general.cfg |
crystallising proteins from dilute solution is in general rather difficult , yet the crystalline form is required in order to perform x - ray crystallography @xcite .
x - ray crystallography is currently the best method of determining the all - important structure of a protein in its native state .
this has lead to considerable interest in how proteins crystallise ; see ref . for a recent review from a colloidal physics perspective and refs . for recent work . in a dilute solution
there may be a large free energy barrier to nucleation of the crystalline phase @xcite , which prevents nucleation on accessible time scales . here
, we consider a generic model protein which in the _ absence _ of polymer already has coexistence between a dilute solution and a crystalline phase .
we determine how the barrier to homogeneous nucleation of the crystalline phase changes when non - adsorbing polymer coils which are larger than the protein molecules are added to the protein solution .
we find that the polymer lowers the nucleation barrier .
this is to be expected as it is well - known that non - adsorbing polymer induces an effective attraction between the protein molecules , and an attraction will favour the dense crystalline phase .
however , what is much less common , is that for large polymer coils it is essential to account for the fact that the polymer - induced attraction is strongly non - pairwise additive . a naive calculation using in the crystalline phase , the effective pair potential in dilute phase of protein molecules , greatly overestimates the effect of the polymer .
protein molecules are typically a few nms across and the root - mean - square end - to - end distance of a flexible polymer can easily be several times this .
the protein
protein interaction consists of a repulsive core of diameter @xmath0 and an attraction which is sufficiently strong to cause the equilibrium behaviour to be coexistence between a dense crystal and a dilute fluid phase . we do not need to specify the potential explicitly as we will require only the fluid - crystal interfacial tension @xmath1 and the difference in chemical potential between that of the crystal and that of the fluid , @xmath2 .
we use the subscript 0 to indicate quantities before polymer is added .
our results should therefore apply to a very wide range of proteins despite potentially very large variations in the structure and interactions of different proteins @xcite .
the polymer is non - adsorbing : the interaction between the protein molecule and a monomer is purely repulsive .
there has been extensive work done on systems of small colloidal particles immersed in polymer solutions of larger polymer coils which do not adsorb on the particles @xcite .
scaling theory for these systems was pioneered by de gennes @xcite , see also refs . .
field - theoretic methods have yielded results which are exact for small hard particles and ideal polymers @xcite , as well as approximate results for polymers in good solvents @xcite .
work has also been done using integral equations , see ref . and references therein .
it is well understood that non - adsorbing polymer increases the chemical potential of particles and induces an attraction between these particles which is not pairwise additive : the interaction of three close - by protein molecules is not the sum of three pair potentials @xcite . however , most of this work ( exceptions are refs . ) only considered a few particles , two or three , whereas a nucleus of a crystalline phase contains 10s of tightly - packed particles . because of this , the deviations from pairwise additivity observed here are much stronger than found in previous work .
the next section briefly outlines classical nucleation theory for the solution before polymer is added .
section 3 then describes the effect of polymer and section 4 is a conclusion .
the interaction between protein molecules is taken to include a steeply repulsive core and an attraction .
this attraction may be isotropic or highly anisotropic or a mixture of the two , if isotropic it may or may not be short - ranged ; all we require is that it render a dilute solution of the protein metastable with respect to a crystalline phase of the protein . by metastable
we mean that the fluid phase is not the equilibrium phase , it does not correspond to the absolute minimum in the appropriate free energy , but the fluid phase is dynamically stable with respect to crystallisation for times much larger than the characteristic time of the solution .
see ref . for a general introduction to the properties and behaviour of metastable fluids .
classical nucleation theory @xcite assumes that the nucleus of the crystalline phase has a free energy @xmath3 which is the sum of a bulk term and surface term .
see the book of debenedetti @xcite for an excellent introduction to classical nucleation theory .
the protein molecules are modeled by spherical particles of diameter @xmath0 .
the bulk term is equal to the number of molecules in the nucleus , @xmath4 , times the chemical potential difference @xmath5 , where @xmath6 is the chemical potential of the crystalline phase and @xmath7 is the chemical potential of the fluid phase which contains the nucleus .
the fluid phase we are considering is not the true equilibrium phase , nucleation occurs from a fluid which is at a higher chemical potential than the crystal .
thus @xmath8 is negative .
the surface term is the surface area of the nucleus times the surface tension @xmath9 of the bulk interface between the coexisting crystalline and fluid phases .
the surface area of the nucleus is obtained by assuming that the nucleus is a sphere of radius @xmath10 which is related to @xmath4 by assuming that the density of spheres within the nucleus is equal to the bulk density of the crystalline phase , taken to be @xmath11 : the density of a close - packed simple cubic lattice
. then @xmath12 .
putting this all together , @xmath13 the first term in @xmath3 is the bulk term , which is negative and decreases linearly with @xmath4 , and the second term is the surface term which is positive and increases as @xmath14 .
thus @xmath3 passes through a maximum , denoted by @xmath15 , at some value of @xmath10 , denoted by @xmath16 .
this maximum is the top of the free energy barrier to nucleation .
the nucleus with radius @xmath16 and @xmath17 spheres is called the critical nucleus . taking the derivative of eq .
( [ df ] ) with respect to @xmath10 and equating to zero , @xmath18 and inserting this value of @xmath4 in eq .
( [ df ] ) @xmath19 nucleation is an activated process with @xmath15 as the barrier height and so the nucleation rate varies as @xmath20 , where @xmath21 and @xmath22 are boltzmann s constant and the temperature @xmath22 , respectively . to increase the rate of nucleation of the crystalline phase of the protein we need to reduce the barrier @xmath15 . in the absence of polymer @xmath15 and @xmath16
are obtained from eqs .
( [ ncnt ] ) and ( [ cnt ] ) by setting @xmath23 and @xmath24 .
now we consider a solution of both protein molecules and flexible polymer molecules . the interaction between a protein molecule and a monomer
is taken to be strictly repulsive .
figure [ figschem ] is a schematic of the nucleus in a solution of polymer .
we remark that this simple model is not adequate for all polymers , for example it is not adequate for peg ( poly(ethylene glycol ) ) .
we consider the case where the polymer molecules are larger than the protein molecules and the polymer solution is semidilute . as the protein molecules are only a few nms in diameter this regime is easily accessible in experiment . in the other limit , where the polymer s radius is smaller than that of the colloidal particles ,
the polymer induces a short - ranged depletion attraction between the particles .
such attractions are to a good approximation pairwise additive @xcite and therefore straightforward to treat
. see refs . for work where the polymer is at most as large as the colloid .
see the book of de gennes @xcite for an introduction to semidilute polymer solutions .
they are characterised by a correlation length @xmath25 which scales as @xmath26 , where @xmath27 is the monomer concentration . the osmotic pressure exerted by the polymer @xmath28 .
the crystalline phase of the protein is dense , the gaps between protein molecules available to polymer are around @xmath29 or less across . forcing a polymer molecule into a tube of diameter @xmath29 costs roughly @xmath30 in free energy per @xmath29 of the tube s length as the polymer collides with the walls of the tube at intervals of roughly its diameter .
thus , the free energy density inside such a crystal , either a bulk crystal or a crystalline nucleus , will be of order @xmath30 per @xmath31 volume whereas in solution it is @xmath30 per @xmath32 volume .
for @xmath25 larger than @xmath0 , which we assume throughout , the free energy density inside a crystal is so much higher than outside that to a good approximation there is no polymer inside the crystal .
this agrees with calculations using the theory of ref . .
so , as the crystalline nucleus is impermeable to polymer it interacts with the polymer , as a hard particle of radius @xmath10 .
the interaction @xmath33 of a hard spherical particle of diameter @xmath0 with a semidilute solution of polymer in a good solvent with a correlation length @xmath25 is @xcite @xmath34 where we took the exponent @xmath35 to have its flory value of @xmath36 @xcite .
@xmath33 is the work done (= the difference in excess chemical potential ) in taking a particle from a pure solvent and inserting it into the polymer solution .
this assumes that the interaction between a protein and the monomers of a polymer is purely repulsive and that there is not much more than one protein molecule per @xmath32 volume .
the notation @xmath37 indicates that we have neglected a coefficient of order unity ; here we derive only the scaling behaviour of the interactions with respect to the relevant length scales .
equation ( [ wsmall ] ) can be derived in a couple of ways , see refs . .
one way is to note that when @xmath38 the interaction between a particle and polymer must scale as the density of monomers , which scales as @xmath39 when @xmath40 . as @xmath41 is dimensionless
, we require that @xmath39 appears as the ratio @xmath42 which gives eq .
( [ wsmall ] ) .
straightaway we can derive an approximation for the contribution of polymer to the difference in excess chemical potential @xmath8 between the dense crystalline phase of the protein , and the semidilute polymer solution . it is @xmath43 where the second term on the right hand side is the change in free energy of the polymer when a single protein molecule is removed from the semidilute polymer solution and the last term is the change in free energy of this solution when it is compressed by the crystalline phase of the protein expanding by the volume of one protein .
the osmotic pressure of a polymer solution @xmath44 is related to @xmath25 by @xmath28 .
thus , the last term is equal to @xmath45 and so is smaller than the second term , indeed it is of order terms we have dropped and so we drop it and obtain our final expression for @xmath46 in the presence of polymer , @xmath47 for a large nucleus , radius @xmath48 , the width of the nucleus polymer - solution interface , @xmath25 , is small and so the capillarity approximation which underlies classical nucleation theory still holds : @xmath3 is still a sum of bulk and interfacial terms .
the bulk term is given by the @xmath8 of eq .
( [ dmup ] ) times the number of protein molecules in the nucleus . in the presence of semidilute polymer the the interfacial tension of a flat interface , @xmath9 is @xcite , @xmath49 where @xmath1 is the interfacial tension in the absence of polymer .
bearing in mind that @xmath1 will be at least of order @xmath50 we see that the fractional modification to @xmath1 due to polymer is of order @xmath51 . for a @xmath2 of order @xmath30 , the fractional change to @xmath8 due to polymer
is , see eq .
( [ dmup ] ) , of order @xmath42 . within our crude scaling theory
we only consider leading order contributions and so neglect the contribution of the polymer to @xmath9 , leaving the only contribution of polymer as the change in @xmath8 given by eq .
( [ dmup ] ) . in the opposite limit for @xmath10 , @xmath52 ,
then the contribution of polymer to the free energy of the nucleus is no longer the sum of bulk and interfacial terms , i.e. , it is not the sum of terms scaling as @xmath53 and as @xmath54 . for
@xmath55 the contribution of the polymer to the free energy of formation of the nucleus @xmath15 is the sum of two terms : the first is the number of protein molecules in the nucleus times minus the increase in chemical potential in solution due to polymer , eq .
( [ wsmall ] ) , and the second is the free energy cost of inserting the nucleus into the polymer solution .
this second term is the free energy cost due to the nucleus excluding polymer from a sphere of radius @xmath10 . because the polymer can not penetrate into the nucleus it interacts with the polymer as a single particle , and so this free energy cost is just eq .
( [ wsmall ] ) with @xmath0 replaced by @xmath56 . adding the two terms together we have the contribution of polymer to @xmath3 , @xmath57 where @xmath58 is the free energy of formation of the nucleus before polymer is added . for @xmath10 a few times @xmath0 and @xmath59 the second term in eq .
( [ smallr ] ) dominates the last term .
so , we neglect the last term in eq .
( [ smallr ] ) .
then as the remaining term is linear in @xmath4 it is in effect a shift in @xmath8 and indeed is the same shift in @xmath8 we found in the opposite limit , @xmath48 .
thus we conclude that as for @xmath48 the leading order contribution of semidilute polymer to the nucleation barrier is simply to shift @xmath8 by the amount given in eq .
( [ dmup ] ) .
so , we have shown that in both limits , @xmath48 and @xmath55 , the effect of adding polymer in the semidilute regime is to shift the chemical potential difference @xmath8 by an amount of order @xmath60 , which will be a few tenths of @xmath30 .
having shown that it holds in both limits we assume that it also holds for @xmath61 as well .
the shift varies as @xmath39 and so increases linearly with the density of polymer .
the size of the critical nucleus also decreases , eq .
( [ ncnt ] ) .
thus , if decreasing @xmath8 by a few tenths of @xmath30 is enough to reduce the nucleation barrier sufficiently to make the nucleation rate significant then adding semidilute polymer is a viable way of inducing nucleation of the crystalline phase of small colloidal particles such as proteins and micelles .
a pair of particles in a semidilute polymer solution separated by a distance @xmath62 of order @xmath25 or less feel a polymer - induced attraction towards each other , which we denote by @xmath63 @xcite . for small particles , @xmath64 , and for separations @xmath65 much smaller than the correlation length @xcite @xmath66
naively it might be thought that the contribution of polymer to the free energy change @xmath46 could be estimated using this pair attraction plus a mean - field approximation , which would correspond to the approximation @xmath67 which is just the integral over the number density , @xmath11 in the crystalline phase , times the potential . as usual with a mean - field approximation , correlations in the positions of particles are neglected and one protein is considered to interact with surrounding particles which are at the mean density .
the above integral is @xmath11 times the contribution of polymer to the second virial coefficient .
see eisenriegler @xcite for a more accurate treatment of the contribution of polymer to the second virial coefficient .
so , inserting eq .
( [ hanke ] ) , into eq .
( [ mf1 ] ) , we obtain @xmath68 if we compare eqs .
( [ dmup ] ) and ( [ mf ] ) we see that the assumption of a pairwise additive potential plus the mean - field approximation predicts the wrong scaling with @xmath69 of the polymer contribution to @xmath46 .
it greatly overestimates , by a factor of order @xmath70 , the effect of adding polymer whose correlation length is large with respect to the protein diameter . naively using the pair potential between an isolated pair of particles , ( a pair of particles with no others within a distance @xmath25 of them ) , as the pair interaction within a dense
, number density @xmath71 ( here the crystal ) , is qualitatively wrong .
we have estimated the change in the barrier to nucleation when polymer is added to a metastable dilute protein solution .
the polymer is semidilute , and has a correlation length @xmath25 larger than the diameter of the protein @xmath0 ; the polymer does not adsorb onto the surface of the protein .
when polymer is added , the barrier @xmath15 is reduced due to the increase in the chemical potential of the protein in solution , possibly by enough to allow nucleation .
however , the effect is not large , @xmath8 decreases by only a few tenths of @xmath30 .
also , a naive use of the polymer induced attraction between a pair of proteins yields results which are qualitatively wrong : they grossly overestimate the effect of adding polymer .
this does not mean that an effective potential approach to the effect of polymer can not reliably estimate the effect of large polymer molecules , just that this effective potential is many - body in nature and can not simply be approximated by just a pair potential .
this conclusion also applies to the phase behaviour of a system of hard - sphere - like colloidal particles and larger polymer molecules @xcite .
finally , we note that the second virial coefficient @xmath72 is often used to measure the strength of the attractions in a protein solution .
@xmath72 is given by @xmath73 where @xmath74 is the second virial coefficient in the absence of polymer and we have assumed not only that @xmath75 is small but that @xmath74 is dominated by interactions with a range @xmath76 : then the second virial coefficient is the sum of the two terms in eq .
( [ b2 ] ) .
so as the polymer density _ increases _ , @xmath25 decreases and the contribution of polymer to @xmath72 _ decreases_. but from eq .
( [ dmup ] ) we see that as the polymer density increases it has a larger and larger effect on the nucleation barrier .
so , the effect of polymer on @xmath72 is _
anti_-correlated with its effect on @xmath15 , i.e. , when one goes up the other goes down , not correlated as we might naively have expected .
as emphasised by evans and coworkers , see for example refs .
, the interaction is only strictly pairwise additive if the polymer is both ideal and does not interact with a particle more than @xmath77 times its radius away . however , if the polymer is larger than this but still small with respect to the protein molecule , the potential will be close to pairwise additive .
then the simultaneous interaction of a polymer molecule with more than two protein molecules ( which is the origin of deviations from pairwise additivity ) will be weak .
in the other limit , which we consider here , where the polymer is much larger than the protein , the polymer molecule can potentially interact with _ many _ protein molecules , inducing an effective interaction which is very far from being pairwise additive . | starting from a protein solution which is metastable with respect to the crystalline phase , the effect of adding semidilute non - adsorbing polymer is considered .
it is found to increase the chemical potential of the protein by a few tenths of kt , which may be enough to lower the barrier to nucleation of the crystalline phase by enough to allow crystallisation .
it is also shown that assuming that the polymer induces a pairwise additive attraction leads to qualitatively incorrect results . 2 |
in this talk i want to discuss some results related to saturation and multiple collisions obtained in collaboration with emil avsar , christoffer flensburg , and leif lnnblad @xcite . in high energy @xmath0 collisions
the minijet cross section is much larger than the total cross section .
this implies that each event contains on average more than one minijet pair , and multiple subcollisions are an essential feature of high energy hadronic reactions . a formulation in transverse coordinate space and the eikonal approximation are particularly suited for a treatment of these features , and has been successfully applied to @xmath1 collisions , both for total and diffractive scattering cross sections @xcite .
the application of the eikonal formalism is , however , mainly applicable for inclusive observables , and a description of exclusive final states is more easy in momentum space . at present the most successful model for high energy @xmath0 collisions
is the pythia model by sjstrand and coworkers @xcite , which has been extensively tuned to tevatron data by r. field @xcite .
there are , however , a set of problems connected to these tunes .
how does the partonic final state hadronize ?
a fit to data needs strong colour reconnections , for which we have a poor theoretical understanding .
also it is difficult to properly include effects of correlations and fluctuations in momentum space cascades . in this talk
i want to discuss a new approach based on mueller s dipole cascade model , which is an attempt to combine the good features of formulations in transverse coordinate space and in momentum space .
at 14 tev , from ref . @xcite . ] in collinear factorization the inclusive parton scattering cross section diverges like @xmath2 for small @xmath3 .
an estimate of the integrated jet cross section above a cut @xmath4 is shown in fig .
[ fig : ptint ] for @xmath0 scattering at 14 tev .
we here see that for @xmath5 gev the integrated jet cross section exceeds the total cross section , which implies that an average collision must have several hard subcollisions .
multiple collisions have also been directly observed in collider experiments .
thus four - jet events with pairwise back - to - back jets have been seen by afs at the isr , by ua2 at the cern s@xmath6s collider , and by cdf at the tevatron . at the tevatron
the most clear signal is observed in the analysis of 3 jets + a prompt photon @xcite .
it was early suggested that the increase in the @xmath0 cross section is driven by minijet production , and that minimum bias events are dominated by ( semi)hard parton subcollisions @xcite .
this is also an essential assumption in the pythia model .
this model is also based on collinear factorization , which implies that a cutoff is needed for the singularity at small @xmath3 .
this is motivated by the fact that hadrons are colour neutral , and therefore the coulomb potential must be screened for large impact parameters or small @xmath3 .
fits to collider data give a cutoff at around 2 gev , slowly increasing at higher energies .
( a similar cutoff is also obtained naturally in the @xmath7-factorization formalism @xcite . in the dipole cascade model the transverse momentum @xmath7 of a colliding gluon is related to the dipole size in transverse coordinate space , and therefore also to the screening length . ) for the more recent version of pythia this cutoff gives typically 2 - 3 interactions per event at the tevatron , and 4 - 5 at the lhc .
an important feature is also that the subcollisions are correlated .
central collisions have many interactions , while peripheral collisions have few . in the experimental analyses
this is described in terms of an effective cross section @xmath8 .
the cross section @xmath9 for the simultaneous hard reactions @xmath10 and @xmath11 ( with @xmath12 ) is written as @xmath13 .
if the interactions @xmath10 and @xmath11 were uncorrelated we would have @xmath8 equal to the inelastic nondiffractive cross section ( @xmath14 50 mb at the tevatron ) .
instead the experimental results give the much smaller result @xmath15 mb , which thus corresponds to a very strong enhancement .
although the pythia model has been tuned to fit most of the cdf data on minimum bias and underlying events , there are a number of open questions : - how does a many - parton system hadronize in an event with multiple hard subcollisions ?
the result expected in the string hadronization model is illustrated in fig .
[ fig : recoupl ] . in a hard gluon - gluon subcollision the outgoing gluons will be colour - connected to the projectile and target remnants , as shown in fig .
[ fig : recoupl]a .
initial state radiation may give extra gluon `` kinks '' , which are ordered in rapidity . a second hard scattering
would naively be expected to give two new strings connected to the remnants as in fig .
[ fig : recoupl]b . as a result
this would give almost doubled multiplicity .
this is not in agreement with data . in the successful fits @xcite it
is instead assumed that the gluons are colour reconnected , so that the total string length becomes as short as possible , see fig . [
fig : recoupl]c .
this colour reconnection implies that a minimum number of hadrons share the transverse momentum of the partons . - in most analysis , including pythia , it is assumed that the impact parameter and momentum distributions of the interacting partons factorize . in a dipole cascade model
strong non - factorizing correlations are expected between impact parameter , momentum , and multiplicity of the colliding partons in a proton .
besides such correlations , also the expected large fluctuations in multiplicity and impact parameter distributions are more difficult to include in a momentum space formalism .
a formalism in transverse coordinate space is very suitable for describing rescattering and multiple collisions . in a process where a particle undergoes successive interactions with transverse momenta @xmath16 , the resulting transverse momentum @xmath17
is given by a convolution of the different interactions .
as the fourier transform of a convolution is given by a simple product , we see that in impact parameter space the multiple interactions are described by a product of the @xmath18-matrix elements for the individual interactions : @xmath19 thus for @xmath20 we find @xmath21 .
a coulomb field which is boosted is contracted to a flat pancake , with a dominantly transverse electric field @xmath22 and an orthogonal transverse magnetic field with the same magnitude . here
@xmath23 is the ( two - dimensional ) distance between the position of the central charge and the point of observation ( see fig . [
fig : ww ] ) . the pulse will be very short in time , and can be approximated by a @xmath24-function : @xmath25 the frequency distribution is given by the fourier transform of @xmath26 , which is a constant .
thus the distribution of photons , or gluons , seen by an observer at point @xmath23 is given by @xmath27 inside a proton there is a very complicated colour field .
we may however imagine that within some distance @xmath28 from a colour charge it is approximately a coulomb field , while for larger distances the charge is screened by other charges .
if @xmath28 is around 0.1 fm this would give an effective cutoff for hard subcollisions with @xmath29 gev , as obtained in the fits to experimental data . a bremsstrahlung gluon from this field
will change the colour of the initial charge .
if the initial charge is a red quark , it may emit a red - antiblue gluon and change its colour to blue .
the result is that the initial red coulomb field now terminates at the gluon charge , and a blue - antiblue dipole field is formed between the quark and the gluon .
the emission of softer gluons now get two separate contributions , one from the modified coulomb field and one from the new colour dipole between the quark and the gluon .
the repeated emission of more gluons results in a cascade , as discussed in the next section . a dipole cascade model in transverse coordinate space was developed by a. mueller in a series of papers @xcite .
we study an initial colour neutral quark - antiquark system .
the colour dipole field gets two contributions of the form in eq .
( [ eq : efield ] ) , as shown in fig .
[ fig : dipolesplit]a . adding these with opposite signs the resulting transverse colour - electric field @xmath30 is given by @xmath31 where @xmath32 is the size of the parent dipole , and @xmath33 and @xmath34 are the distances from the point @xmath23 to the two charges ( see fig . [
fig : dipolesplit]a ) . as discussed above this
implies that the probability to emit a gluon in the point @xmath23 becomes @xmath35 if the initial charges were red and antired , the emitted gluon may be red - antiblue , changing the initially red charge to blue .
therefore such an emission implies that the dipole is split in two connected dipoles , one blue - antiblue and one red - antired .
these can then also split repeatedly in a dipole cascade , as shown in fig .
[ fig : dipolesplit]b . in ref .
@xcite it is demonstrated that this cascade reproduces the ll bfkl evolution .
when two such cascades collide as in fig .
[ fig : dipcoll ] , two dipoles ( @xmath36 and @xmath37 with endpoints ( @xmath38 ) and ( @xmath39 ) respectively ) may interact via gluon exchange .
adding coherently the exchange between charge or anticharge in both dipoles gives the interaction probability @xmath40 as the exchanged gluon carries colour , the dipole chains are reconnected , as also indicated in fig .
[ fig : dipcoll ] .
the result is two dipole chains connecting the remnants of the projectile and the target systems , as also illustrated in fig .
[ fig : recoupl ] .
it is also possible that several pairs ( @xmath41 ) interact . in the eikonal approximation the total , diffractive , and elastic cross sections
are given by @xmath42 we note that as the parentheses always are smaller than unity , unitarity is always satisfied .
we also note that diffractive excitation , which is given by the difference between the second and third lines in eq .
( [ eq : eikonalcrossections ] ) , is directly determined by the fluctuations in the cascade evolution . a schematic picture of a collision between two evolved dipole cascades is shown in fig .
[ fig : finalstate ] . in this example
there are in the cms ( indicated by a dashed line ) three separate subcollisions .
these also correspond to the exchange of three pomerons .
they result in two closed loops formed by chains of colour dipoles , in addition to the two dipole chains which connect the projectile and target remnants .
[ fig : finalstate ] also contains a dipole loop ( denoted @xmath10 ) inside the evolution of the left system , before it collides .
such loops are , however , not included in mueller s model , which only accounts for loops cut in the specific lorentz frame used for the calculation .
this implies that this model is not lorentz frame independent . ) within the left system , caused by a `` dipole swing '' .
the parts of the evolutions marked @xmath11 and @xmath43 do not interact , and should be treated as virtual . ]
although mueller s cascade model , and also the balitsky - kovchegov evolution equation , reproduce ll bfkl evolution and satisfies the unitarity constraints , they have a set of problems : * ll bfkl is not good enough .
nll corrections are very large . *
non - linear effects in the evolution are not included .
* massless gluon exchange implies a violation of froissart s bound .
* it is difficult to include fluctuations and correlations ; the bk equation represents a mean field approximation .
* they can only describe inclusive features , and not the production of exclusive final states . *
analytic calculations are mainly applicable at extreme energies , well beyond what can be reached experimentally .
non - leading effects , effects of energy conservation and a running coupling , are often easier included in monte carlo simulations .
a simulation of muller s cascade model has been presented by salam @xcite . however , as the dipole splitting diverges for small dipole sizes , @xmath44 , ( see eq .
( [ eq : dipolesplit ] ) ) this simulation is hampered by numerical difficulties .
it is noticed that the fluctuations in the number of dipoles is very large , which causes serious problems .
in refs . @xcite we describe a modification of mueller s cascade model with the following features : * it includes essential nll bfkl effects . *
it includes non - linear effects in the evolution . *
it includes essential correlations and fluctuations .
* it can also describe exclusive final states .
the model is also implemented in a monte carlo simulation program . here
the nll effects significantly reduce the production of small dipoles , and thereby also the associated numerical difficulties with very large dipole multiplicities are avoided .
confinement corrections are also included , but this is numerically less important than nll and non - linear effects .
the nll corrections to bfkl evolution have three major sources @xcite : * _ the running coupling . _ + this is relatively easily included in a mc simulation process . *
_ non - singular terms in the splitting function . _
+ these terms suppress large @xmath45-values in the individual branchings , and prevent the daughter from being faster than her recoiling parent .
most of this effect is taken care of by including energy - momentum conservation in the evolution .
* _ projectile - target symmetry .
_ + this is also called energy scale terms , and is essentially equivalent to the so called consistency constraint .
this effect is taken into account by conservation of both positive and negative lightcone momentum components , @xmath46 and @xmath47 .
the treatment of these effects includes effects beyond nll , in a way similar to the treatment by salam in ref .
thus the power @xmath48 , determining the growth for small @xmath49 , is not negative for large values of @xmath50 .
as mentioned above , dipole loops within the evolution , as indicated by the letter @xmath10 in fig .
[ fig : finalstate ] , are not included in mueller s cascade model or in the jimwlk or bk equations . like for dipole scattering the probability for such loops
is given by @xmath50 , and therefore formally colour suppressed compared to dipole splitting , which is proportional to @xmath51 .
these loops are therefore related to the probability that two dipoles have the same colour .
two dipoles with the same colour form a quadrupole .
such a field may be better approximated by two dipoles formed by the closest colour - anticolour charges .
this corresponds to a recoupling of the colour dipole chains , as indicated in fig .
[ fig : swing ] .
we call this process a dipole `` swing '' .
although this mechanism does not give an explicitly frame independent result , mc simulations show that it is a very good approximation ( see also fig .
[ fig : sigmatot ] below ) .
we note that in this formulation the number of dipoles is not reduced .
the saturation effect is obtained because the swing favours the formation of smaller dipoles , which have a smaller cross section .
thus in an evolution in momentum space the swing would not correspond to an absorption of gluons below the saturation line @xmath52 ; it would rather correspond to lifting the gluons to higher @xmath7 above this line .
the exchange of massless gluons gives an interaction of infinite range , which eventually will violate froissart s bound .
this long range interaction is suppressed by introducing an effective gluon mass .
thus the gluon propagater @xmath53 is replaced by @xmath54 .
this implies that the expression for the transverse electric field in eq .
( [ eq : efield ] ) should be replaced by @xmath55 , and in the interaction probability in eq .
( [ eq : intprob ] ) the logarithms @xmath56 are replaced by @xmath57 . here
@xmath58 and @xmath59 and @xmath60 are modified bessel functions . for small values of @xmath44
the expressions are unchanged , but when @xmath44 becomes larger than @xmath61 they become exponentially suppressed . as the size of a dipole
is also associated with ( the inverse of ) its transverse momentum , the event depicted in fig .
[ fig : finalstate ] can also be interpreted as an exclusive final state with definite parton momenta .
it is true that taking the fourier transform is not exactly the same as replacing @xmath23 by @xmath62 .
therefore , when there is a conflict , what we really are generating is perhaps not @xmath44 with the approximation @xmath63 , but rather generating @xmath7 with @xmath64 .
there are also some other problems which have to be solved . in the evolution of a particle with a high value for the lightcone momentum , @xmath46 , but very little @xmath47
, the gluons in the cascade will accumulate a large deficit of @xmath47 , which has to be `` payed back '' when it collides with a target having a large @xmath47 but small @xmath46 .
dipole chains like the ones indicated by @xmath11 and @xmath43 in fig .
[ fig : finalstate ] , which do not interact , can not come on shell and have to be treated as virtual .
thus only gluons which are connected to daughters involved in the collisions should be included in the initial state radiation .
an important feature is that all dipoles are connected in chains , which implies strong non - trivial correlations @xcite .
a toy model calculation by r. corke @xcite shows that such correlations can have a big effect on the need for colour reconnections in the final state .
the full calculation of the final state properties is not yet implemented in the mc , but will be ready in the near future .
for the results presented in this section we have used a simple model for the proton wave function , in form of three dipoles in an equilateral triangle . for larger @xmath65 the wavefunction for a virtual photon
is determined by perturbation theory , but for smaller @xmath65 it is necessary to include a nonperturbative component representing the soft hadronic part of the photon . for more details
see ref .
@xcite . scattering cross section .
results are presented for evolution with and without the dipole swing mechanism . the one pomeron result and
the result obtained in a frame where one of the protons is almost at rest is also shown .
_ right _ : @xmath1 total cross section compared to data from h1 @xcite and zeus @xcite .
the result is plotted as a function of the gbw scaling variable @xmath66 .
, title="fig : " ] scattering cross section .
results are presented for evolution with and without the dipole swing mechanism . the one pomeron result and
the result obtained in a frame where one of the protons is almost at rest is also shown .
_ right _ :
total cross section compared to data from h1 @xcite and zeus @xcite .
the result is plotted as a function of the gbw scaling variable @xmath66 .
, title="fig : " ] the total @xmath0 cross section is shown in fig .
[ fig : sigmatot]a .
we note that the one - pomeron result , which neglects unitarity constraints , is about a factor four too high at tevatron energies .
also saturation effects inside the evolution , simulated by the swing , have a significant effect , reducing the cross section by about 30% .
we also see that including the swing makes the result obtained in the rest frame of the target ( long dashed line ) very close to the result in the cms ( solid line ) , which shows that the result is almost lorentz frame independent .
the total @xmath1 cross section is presented in fig .
[ fig : sigmatot]b against the golec - biernat wsthoff scaling parameter @xmath67 , where @xmath68 with @xmath69 , @xmath70 , @xmath71 @xcite .
we note that our result , in agreement with the data , satisfies geometric scaling both for small @xmath72 , where the result is influenced by saturation , and for larger @xmath72 where saturation is not essential . elastic scattering , together with data from ua4 @xcite and cdf @xcite .
the figure also includes a prediction for the lhc . ]
[ fig : pptdep ] shows the elastic @xmath0 cross section . as the real part of the amplitude
is neglected the diffractive dip is a real zero .
the position of this dip can be tuned at one energy , but the change to smaller @xmath73 at higher energies is a result of the evolution , which can not be tuned . for @xmath74 = 82 gev as function of @xmath65 , compared to data from h1 @xcite .
_ right _ : the cross section for @xmath75 for @xmath65 = 4.8 , 13.5 , and 32 @xmath76 as function of @xmath74 , compared to data from h1 @xcite and zeus @xcite .
two different wavefunctions have been used for the @xmath77-meson , boosted gaussian ( solid line ) and dgkp ( dashed line).,title="fig : " ] for @xmath74 = 82 gev as function of @xmath65 , compared to data from h1 @xcite .
_ right _ : the cross section for @xmath75 for @xmath65 = 4.8 , 13.5 , and 32 @xmath76 as function of @xmath74 , compared to data from h1 @xcite and zeus @xcite .
two different wavefunctions have been used for the @xmath77-meson , boosted gaussian ( solid line ) and dgkp ( dashed line).,title="fig : " ] quasielastic @xmath1 scattering is also well reproduced by the model , including the dependence on @xmath65 , @xmath74 , and @xmath73 . as two examples fig .
[ fig : quasi ] shows the @xmath65-dependence for dvcs and the @xmath74-dependence for @xmath75 . for the latter case two
different @xmath77-meson wavefunctions were used , a boosted gaussian wavefunction @xcite and the dgkp @xcite wavefunction .
both models show similar growth with @xmath74 , but our normalization agrees better with the boosted gaussian wavefunction .
as discussed above diffractive excitation is directly determined by the fluctuations in the evolution . as an example , if we calculate the expression @xmath78 in a frame where the right - moving proton is evolved @xmath79 rapidity units , it gives the cross section for single diffractive excitation of this proton to masses satisfying @xmath80 , with @xmath81 gev . here
@xmath82 and @xmath32 indicate averages over left- and right - moving proton cascades respectively .
the result obtained when varying @xmath79 is shown in fig .
[ fig : diff]a compared with corresponding results from the tevatron .
fm at two different energies , @xmath83 9 and 11 . our results ( solid and dashed lines )
are compared to those from the kowalski - teaney @xcite model ( dot - dashed and dotted lines),title="fig : " ] fm at two different energies , @xmath83 9 and 11 .
our results ( solid and dashed lines ) are compared to those from the kowalski - teaney @xcite model ( dot - dashed and dotted lines),title="fig : " ] we also note that as the fluctuations in the evolution are quite large , less fluctuations are needed in the impact parameter profile .
thus this profile is more gray , and less black and white , than in fits where the fluctuations in the evolution are not taken into account .
this is seen in fig .
[ fig : diff]b , which shows a comparison between our model and the result by kowalski and teaney @xcite for the scattering of a dipole against a proton . in the latter analysis
the fluctuations in the evolution are not included .
* a new dipole formulation of high energy collisions in transverse coordinate space is presented * it has the following main ingredients : * * nll corrections to bfkl * * non - linear effects : saturation and multiple subcollisions * * confinement effects * * includes momentum - impact parameter correlations * * simple proton and photon models * * mc implementation * it gives a fair description of data for : * * total cross sections for @xmath0 and @xmath1 collisions * * ( quasi-)elastic scattering in @xmath0 and @xmath1 * * diffractive excitation * to come soon : generation of exclusive final states * wanted : better understanding of the connection to the @xmath73-channel picture of pomeron loops | _ dedicated to the memory of jan kwieciski
_ in high energy collisions saturation and multiple collisions are most easily accounted for in transverse coordinate space , while analyses in momentum space have been more suitable for calculating properties of exclusive final states . in this talk
i describe an extension of mueller s dipole cascade model , which attempts to combine the good features of both these descriptions .
besides saturation it also includes effects of correlations and fluctuations , which have been difficult to account for in previous approaches .
the model reproduces successfully total , elastic , and diffractive cross sections in @xmath0 collisions and dis , and a description of final states will be ready soon . |
the human brain contains about @xmath0 neurons @xcite , and each neuron is connected to approximately @xmath1 other neurons @xcite .
these connections called synapses , are arranged in a highly complex network .
they are responsible for neuronal communication and can be classified into two categories : electrical and chemical synapses @xcite . in electrical synapses the transmission of information from one neuron to another is directly performed from the pre - synaptic cell to the post - synaptic cell via gap junctions . in chemical synapses ,
the process occurs via neurotransmitters , which cross the synaptic cleft and bind to receptors on the membrane of the synaptic cell @xcite .
neurotransmitters may increase or decrease the probability of an action potential of a post - synaptic neuron , and the synapses are called excitatory or inhibitory , respectively @xcite . furthermore , the intensity of the chemical synapses can be modified , in other words , they can be minimised or potentiated @xcite .
the mechanism responsible for these adjustments is known as synaptic plasticity @xcite .
the synaptic plasticity , that is , the ability of synapses to weaken or strenthen over time @xcite is an important property of the mammalian brain .
in addition , the synaptic plasticity is also related to processes of learning and memory @xcite .
this adjustment of the intensities of the chemical synapses can be correlated with phenomena of synchronisation of the neuronal firing @xcite .
the occurrence of synchronisation in some specific areas of the brain may be associated with some diseases , such as the epilepsy and the parkinson s disease @xcite . on the other hand , it is also responsible for some vital brain functions , such as processing of sensory information and motor function @xcite .
methods to suppress synchronisation have been proposed in neuroscience , as the introduction of external perturbations @xcite .
tass and collaborators have verified the possibility of desynchronisation in hippocampal neuronal populations through coordinated reset stimulation @xcite .
meanwhile , popovych and collaborators have found that the introduction of a perturbation in a globally connected neuronal network combined with synaptic plasticity can provide a positive contribution to the firing synchronisation @xcite . in this work ,
we study firing synchronisation in a random hodgkin - huxley neuronal network with plasticity according to spike timing - dependent plasticity ( stdp ) .
this synaptic plasticity model adjusts the connection strengths by means of the temporal interval between pre - synaptic and post - synaptic spikes @xcite .
bi and poo have reported that the change in synaptic efficiency after several repetitions of the experiment is due to the time difference of firing @xcite .
if one pre - synaptic spike precedes a post - synaptic spike , a long - term potentiation occurs , otherwise , a long - term depression appears @xcite .
a computational neuronal network specifies the connection architecture among neurons . a globally coupled hodgkin - huxley neuron model was also considered by popovych and collaborators @xcite .
they studied the synchronisation behaviour considering stdp , and found that the mean synaptic coupling presents a dependence on the input level . in this work
, we consider a random neuronal network with stdp , and input , where the connections are associated with chemical synapses @xcite .
one main result is to show that spike synchonisation in a neuronal network , depending on the probability of connections , can be improved due to spike timing - dependent plasticity .
this improvement is also observed when an external perturbation is applied on the network .
another important result is that the orientation of the connections among neurons with a different spike frequency affect the synchronised behaviour .
this paper is organised as follows : in section ii we introduce the hodgkin - huxley neuronal model . in section iii , we show the random neuronal network . in section iv
, we study the synchronisation considering spike timing - dependent plasticity . finally , in the last section , we draw the conclusions .
one of the most important models in computational neuroscience is the neuronal model proposed by hodgkin and huxley @xcite . in this model ,
the mechanism of generation of an action potential was elucidated in a series of experiments with the squid giant axon .
they found three different ions currents consisting of sodium ( na ) , potassium ( k ) and leak ( l ) mainly due to chlorine .
moreover , there are voltage - dependent channels for sodium , potassium that control the entry and exit of these ions through the cell
. the model is composed of a system of four coupled differential equations given by @xmath2 where @xmath3 is the membrane capacitance ( measured in @xmath4f/@xmath5 ) , @xmath6 is the membrane potential ( measured in mv ) , the function @xmath7 and @xmath8 are the variable of activation for sodium and potassium , and @xmath9 is the variable of inactivation for sodium .
the functions @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath11 are given by @xmath15 the parameters @xmath16 and @xmath17 represent the conductance and reversal potentials for each ion .
the constant @xmath18 is an external current density ( measured in @xmath4a/@xmath5 ) that determines a regime of a single spike ( @xmath19a/@xmath5 ) , or a scheme with periodic spikes ( @xmath20a/@xmath5 ) , as illustrated in fig .
[ fig1](a ) and ( b ) , respectively .
moreover , the spikes frequency increases when the constant @xmath18 increases .
for instance , @xmath20a/@xmath5 and @xmath21a/@xmath5 approximately correspond to @xmath22hz and @xmath23hz , respectively .
the parameters that we use in this work are presented in table [ parametros_hh ] @xcite .
.parameters of the hodgkin - huxley neuronal model with a resting potential equal to @xmath24mv . [ cols="^,^,^",options="header " , ] [ parametros_hh ] a/@xmath5 that shows a single spike with a subsequent resting state , and ( b ) @xmath20a/@xmath5 that presents a regime with periodic firing.,width=302,height=302 ] computational models of neuronal networks depend on the architecture , which specifies how neurons are connected and how the dynamics is applied to each unit or node . in this work ,
we consider a random network , that is , the network is constructed by connecting neurons randomly @xcite .
each connection is included with probability independent from every other connection .
figure [ fig2 ] exhibits a schematic representation of the neuronal network considered in this work .
each neuron is connected to others by randomly chosen neurons with probability @xmath25 .
when @xmath26 we have a global network , where all neurons are connected .
.,width=226,height=226 ] we consider a random neuronal network with chemical synapses where the connections are unidirectional , and the local dynamics is described by the hodgkin - huxley model .
the network is given by @xmath27 where @xmath28 is the membrane potential of neuron @xmath29 ( @xmath30 ) , @xmath31 is a constant current density randomly distributed in the interval @xmath32 $ ] , @xmath33 is the average degree connectivity , and @xmath34 is the coupling strength from the pre - synaptic neuron @xmath35 to the post - synaptic neuron @xmath29 , that is , normally distributed with mean and variance equal to @xmath36 and @xmath37 , respectively @xcite .
we consider an external perturbation @xmath38 , so that each neuron receives an input with a constant intensity @xmath39 during @xmath40ms .
this input is applied with an average time interval around @xmath41ms .
this value is approximately the inter - spike interval of a single neuron .
the neurons are excitatory coupled with a reversal potential @xmath42mv @xcite . the post - synaptic potential @xmath43 is given by @xcite @xmath44 when identical neurons are coupled , the network may exhibit a complete synchronisation among spiking neurons , i.e. , all neurons have identical time evolution of their action potential .
we do not consider identical neurons here , and due to this fact , a complete synchronisation is not possible .
however , a weak synchronisation may be observed .
as diagnostic of spikes synchronisation we use the order parameter given by @xcite @xmath45 where @xmath46 where @xmath47 denotes when a spike @xmath48 ( @xmath49 ) of a neuron @xmath35 occurs ( @xmath50 ) .
the beginning of each spike is considered when @xmath51 .
if the spikes times are uncorrelated , their contribution to the result of the summation is small @xcite .
however , in a globally synchronised state the order parameter magnitude asymptotes the unity .
figure [ fig3](a ) shows the spiking patterns for @xmath52 , and without external perturbation , that is , the neuronal network presents asynchronous dynamics , where the points correspond to spiking neurons , and the absent points correspond to the resting neurons . for @xmath53 , we have globally coupled neurons , and we can see that the network exhibits synchronised spiking , shown in fig . [
fig3](b ) .
the time evolution of the order parameter is plotted in fig .
[ fig3](c ) .
when @xmath52 ( black line ) , the network does not display synchronised behaviour , and as a result the order parameter is typically small with @xmath54 fluctuating around @xmath36 .
however , synchronised behaviour is observed for @xmath53 with order parameter values near unity ( red line ) .
we add an external perturbation ( @xmath38 ) to analyse its effects on the synchronous behaviour .
this way we compute the time averaged magnitude of the order parameter , given by @xmath55 where the values of @xmath56 have been computed by averaging over a temporal length of @xmath57s after discarding a transient of @xmath58s ( @xmath59s and @xmath60s ) . in fig .
[ fig4 ] we can see the time averaged order parameter as a function of the probability for different input amplitudes ( @xmath39 ) .
when the neuronal network has no input , it is possible to observe synchronised behaviour if the probability of connection is large enough ( black circles ) .
however , external inputs are able to desynchronise the spiking neurons , as shown in fig .
[ fig4 ] for @xmath61 ( red triangles ) and @xmath62 ( blue squares ) , where the values of @xmath56 are less than @xmath63 . , ( a ) @xmath52 , and ( b ) @xmath53 . in ( c )
the order parameter is calculated for @xmath52 ( black line ) , and @xmath53 ( red line).,width=302,height=340 ] , for external perturbation with @xmath64 ( black circles ) , @xmath61 ( red triangles ) , and @xmath62 ( blue squares ) .
the bars represent the standard deviation.,width=264,height=226 ]
one of the key principles of behavioural neuroscience is that experience can modify the brain structure , that is known as neuroplasticity @xcite .
although the idea that experience may modify the brain structure can probably be traced back to the 1890s @xcite , it was hebb who made this a central feature in his neuropsychological theory @xcite . with this in mind , we consider spike timing - dependent plasticity according to the hebbian rule . in this plasticity
the coupling strength @xmath34 is adjusted based on the relative timing between the spikes of pre - synaptic and post - synaptic neurons @xcite .
@xmath65 where @xmath66 .
figure [ fig5 ] exhibits the result that is obtained from eq .
( [ eqplast ] ) for @xmath67 , @xmath68 , @xmath69ms , and @xmath70ms . ) as a function of the difference of spike timing of post- and pre - synaptic neuron.,width=264,height=226 ] ( red line ) , @xmath71 ( green line ) , @xmath72 ( blue line ) , and @xmath53 ( black line ) , where it is considered ( a ) @xmath64 , and ( b ) @xmath61.,width=264,height=340 ] the initial synaptic weights @xmath34 are normally distributed with mean and variance equal to 0.1 and 0.02 , respectively .
then , they are updated according to eq .
( [ eqplast ] ) , where @xmath73 . in the absence of an external perturbation
, we can verify by means of fig .
[ fig6](a ) that the averaged synaptic weights can be depressed ( @xmath74 ) or potentiated ( @xmath71 , @xmath72 , and @xmath53 ) depending on the probability of connections . if an input is applied on the neuronal network ( fig .
[ fig6]b ) , the input can have a constructive effect on the synaptic weights ( @xmath72 and @xmath53 ) or destructive effect ( @xmath74 and @xmath71 ) , depending on the probability of connections .
, ( b ) @xmath75 , ( c ) @xmath61 , and ( d ) @xmath62 .
the bars represent the standard deviation.,width=302,height=264 ] the time averaged order parameter in terms of the probability of connection is showed in fig .
[ fig7 ] , considering with ( black circles ) and without ( red triangles ) stdp . in the case for without external perturbation ( fig .
[ fig7]a ) we can see that the values of @xmath56 without stdp are less than with stdp , namely stdp is producing a positive effect on the synchronisation .
increasing the amplitude of the external perturbation , without spike timing - dependent plasticity , the desynchronisation is induced in the neuronal network ( fig .
[ fig4 ] ) . however , considering stdp in the perturbed network is possible to observe alterations to the dynamic behaviour in relation to synchronised states .
figure [ fig7](b ) shows that the stdp enhances the synchronisation for @xmath25 approximately greater than @xmath76 due to a constructive effect on the dynamics of the synaptic weights . on the other hand , for @xmath25 less than @xmath76
the stdp decreases the values of the time averaged order parameter , as a result of depressed synaptic weights . increasing the input intensity ( @xmath61 ) , it is possible to verify synchronisation when the network has neuroplasticity ( fig . [ fig7]c ) .
we can also see an abrupt transition from a desynchronised to a synchronised regime . when the input intensity increases , it is necessary to increase the probability of connection for the stdp counteracts the suppression of synchronisation .
figure [ fig7](d ) exhibits a situation such that the synchronisation is only obtained when the neuronal network presents a global coupling ( @xmath26 ) . with a strong input ,
the stdp does not lead the network to a potentiation of synaptic weights , and this way the synchronisation is suppressed by an external input . ) as a function of the difference of spike timing of post- and pre - synaptic neuron , where the curves are the potentiation ( red line ) and the depression ( black line ) .
( b ) average time difference versus probability of connections for @xmath64 ( red circles ) , @xmath75 ( blue squares ) , @xmath61 ( black triangles ) , and @xmath62 ( green diamonds ) .
( c ) critical probability @xmath77 as a function of the input level @xmath39 .
the linear fit is given by the equation @xmath78.,width=264,height=340 ] the spike synchronisation depends on the probability of connections @xmath25 in a way showing an abrupt transition .
there is a critical point for @xmath79 that can be found by means of the intersection between the curves of potentiation and depression .
figure [ fig8](a ) exhibits the point of intersection with value of @xmath80 approximately equal to @xmath81 .
for @xmath82 the depression ( red line ) of the synaptic strength is larger than the potentiation ( black line ) , while that for @xmath83 the potentiation is larger than the depression . with this value of @xmath80
we can obtain @xmath84 plotting @xmath85 as a function of @xmath25 , as shown in fig . [
fig8](b ) .
it can be seen that the value of @xmath85 decays , and when cross the value @xmath86 we have the critical value of the probability of connections . moreover , when @xmath25 increases not only @xmath85 decreases , but also the variance of the inter - spike intervals decreases .
then , we compute the critical probability as a function of the input level ( fig .
[ fig8]c ) , where we verify a linear increase given by the equation @xmath78 .
figure [ fig7](c ) shows a discontinuous transition between the synchronised and the desynchronised regime .
our aim is to understand how this discontinuous transition appears when the probability of connections is varied .
for this reason , we build a network according to a schematic representation that is showed in fig .
the scheme represents a network of neurons with high ( cyan ball ) and low ( yellow ball ) spiking frequency .
the red arrows represent the connections between neurons from high to low frequency , while the black arrows represent the connections from low to high frequency .
based on the schematic representation that is illustrated in fig .
[ fig9 ] , we consider a neuronal network with @xmath87 , @xmath88 , and @xmath75 .
we separate the neuronal network into @xmath89 neurons with high frequency ( values of @xmath31 within the range @xmath90 $ ] ) and @xmath89 neurons with low frequency ( values of @xmath31 within the range @xmath91 $ ] ) .
figure [ fig10 ] exhibits the time averaged coupling strength as a function of the percentage of connections from neurons with high frequency to neurons with low frequency .
we can see that the time average coupling strength depends on the connections . considering the case for a small percentage of connections from hfn to lfn , the average coupling is small , indicating absence of synchronisation , a situation that changes with increasing the percentage of connections to a synchronised state .
this means that , when the coupling strengths increase , a desynchronised state can suddenly become synchronised .
consequently , the abrupt transition from desynchronised to synchronised state , that is observed in fig . [ fig7 ] , is due to directed synapses among spiking neurons with high and low frequency . and @xmath75.,width=264,height=264 ]
we have been studying a neuronal network model with spiking neurons .
we have chosen , as local dynamics , the hodgkin - huxley model due to the fact that it has essential features of spiking dynamics .
the hodgkin - huxley model is a coupled set of nonlinear differential equations that describes the ionic basis of the action potential .
these equations are able to reproduce biophysical properties of the action potential .
we have used a random coupling architecture where the connections are randomly distributed according to a probability . when the probability is equal to unity we have a globally coupled network .
the connections were considered unidirectional representing excitatory chemical synapses .
we have studied the effects of spike timing - dependent plasticity on the synchronisation in a hodgkin - huxley neuronal network .
studies about spike synchronisation are important to understand not only progressively degenerative neurological disorders , but also processing of the sensory information .
popovych and collaborators @xcite showed that stdp combined with an external perturbation can improve the spike synchronisation in a globally neuronal network .
the novelty in this paper is that we have considered a random neuronal network and we have verified that the spike synchronisation depend on the probability of connections . considering a strong external perturbation the spike synchronisation
is suppressed . however , when there is stdp , depending on the probability of connections , the synchronisation in the perturbed network can be improved due to a constructive effect on the synaptic weights .
we have also shown that the direction of synapses has an important role on the effects of spike timing - dependent plasticity on the synchronisation in a random hodgkin - huxley neuronal network .
this study was possible by partial financial support from the following brazilian government agencies : cnpq , capes and fapesp .
financial support by the spanish ministry of economy and competitivity under project number fis2013 - 40653-p is also acknowledged .
00 chialvo dr .
critical brain networks .
phys a 2004;340:756 - 765 .
gerstner w , kistler w. spiking neuron models : single neurons , populations , plasticity .
cambridge university press : cambridge ; 2002 . bear mf , connors bw , paradiso ma .
neuroscience : exploring the brain .
lippincott williams and wilkins : england ; 2008 .
fundamental neuroscience . academic press : new york ; 2008 .
dayan p , abbott lf .
theoretical neuroscience : computational and mathematical modelling of neural systems .
mit press : massachusetts ; 2001 .
purves d , augustine dj , fitzpatrick d , hall wc , lamantia a - s , mcnamara jo , williams sm . neuroscience .
sinauer associates inc publishers : massachusetts : 2004 .
dagostin aa , mello cv , leo rm . increased bursting glutamatergic neurotransmission in an auditory forebrain area of the zebra finch ( taenopygia guttata ) induced by auditory stimulation . j comp physiol a 2012:198;705 - 716 .
the organization of behavior .
wiley : new york ; 1949 .
citri a , malenka rc .
synaptic plasticity : multiple forms , functions , and mechanisms .
neuropsychopharmacol 2008:33;18 - 41 .
kelso sr , ganong ah , brown th .
hebbian synapses in hippocampus .
proc natl acad sci usa 1986:83;5326 - 5330 .
caporale n , dan y. spike timing - dependent plasticity : a hebbian learning rule .
annu rev neurosci 2008:31;25 - 46 .
zhigulin vp , rabinovich mi , huerta r , abarbanel hd .
robustness and enhancement of neural synchronization by activity - dependent coupling .
phys rev e 2003:67;021901 .
abuhassan k , coyle d , maguire l. compensating for thalamocortical synaptic loss in alzheimers disease .
front comput neurosci 2014:8;65 .
modolo j , bhattacharya b , edwards r , campagnaud j , legros a , beuter a. using a virtual cortical module implementing a neural field model to modulate brain rhythms in parkinsons disease .
front neurosci 2010:4;45 .
hammond c , bergman h , brown p. pathological synchronization in parkinsons disease : networks , models and treatments .
trends neurosci 2007:30;357 - 364 .
nini a , feingold a , slovin h , bergman h. neurons in the globus pallidus do not show correlated activity in the normal monkey , but phase - locked oscillations appear in the mptp model of parkinsonism .
j neurophysiol 1995:74;1800 - 1805 .
uhlhass p , singer w. neural synchrony in brain disorders : relevance for cognitive dysfunctions and pathophysiology .
neuron 2006:52;155 - 168 .
lameu el , batista cas , batista am , iarosz kc , viana rl , lopes sr , kurths j. suppression of bursting synchronization in clustered scale - free ( rich - club ) neuronal networks .
chaos 2012:22;043149 .
popovych ov , tass pa .
desyncrhonizing electrical and sensory coordinated reset neuromodulation .
front hum neurosci 2012:6;58 .
popovych ov , yanchuk s , tass pa .
self - organized noise resistance of oscillatory neural networks with spike timing - dependent plasticity .
sci rep 2013:3;2926 .
nordenfelt a , used j , sanjun maf . bursting frequency versus phase synchronization in time - delayed neuron networks .
phys rev e 2013:87;052903 .
tass pa , silchenko an , hauptmann c , barnikol ub , speckmann e - j .
long - lasting desynchronization in rat hippocampal slice induced by coordinated reset stimulation .
phys rev e 2009:80;011902 .
feldman de .
the spike - timing dependence of plasticity .
neuron 2012:75;556 - 571 .
markram h , gerstner w , sjstrm pj .
spike - timing - dependent plasticity : a comprehensive overview .
front synaptic neurosci 2012:4;8 .
bi gq , poo mm .
synaptic modifications in cultured hippocampal neurons : dependence on spike timing , synaptic strength , and postsynaptic dell type .
j neurosci 1998:18;10464 - 10472 .
gerstner w , sprekeler h , deco g. theory and simulation in neuroscience .
sci 2012:338;60 - 65 .
electron microscopy of synaptic contacts on dendrite spines of the cerebral cortex .
nature 1959:183;1592 - 1593 .
hodgkin al , huxley af . a quantitative description of membrane current and its application to conduction and excitation in nerve .
j physiol 1952:117;500 - 544 .
izhikevich em . which model to use for cortical spiking neurons ?
ieee t neur net 2004:15;1063 - 1070 .
izhikevich em .
dynamical systems in neuroscience : the geometry of excitability and bursting .
mit press : london ; 2006 .
erds p , rnyi a. on random graphs i. publ math 1959:6;290 - 297 .
nordenfelt a , wagemakers a , sanjun maf .
frequency dispersion in the time - delayed kuramoto model .
phys rev e 2014:89;032905 .
nordenfelt a , wagemakers a , sanjun maf .
cyclic motifs as the governing topological factor in time - delayed oscillator networks .
phys rev e 2014:90;052920 .
destexhe a , mainen zf , sejnowki tj . an efficient method for computing synaptic conductances based on a kinetic model of receptor binding .
neural comput 1994:6;14 - 18 .
golomb d , rinzel j. dynamics of globally coupled inhibitory neurons with heterogeneity .
phys rev e 1993:48;4810 .
kuramoto y. international symposium on mathematical problems in theoretical physics .
springer - verlag : new york ; 1975 .
kuramoto y. chemical oscillations , waves , and turbulence .
springer : berlin ; 1984 .
acebrn ja , bonilla ll , vicente cjp , ritort f , spigler r. the kuramoto model : a simple paradigm for synchronization phenomena .
rev mod phys 2005:77;137 - 185 .
ramon y cajal s. degeneration and regeneration of the nervous system .
oxford university press : london ; 1928 .
bliss tvp , gardner - medwin ar .
long - lasting potentiation of synaptic transmission in the dentate area of the unanaestherised rabbit following stimulation of the perforant path .
j physiol 1973:232;357 - 374 .
bliss tvp , collingridge gl . a synaptic model of memory : long - term potentiation in the hippocampus . nature 1993:361;31 - 39 .
. brain mechanisms and learning .
oxford university press : london ; 1961 . | in this paper , we study the effects of spike timing - dependent plasticity on synchronisation in a network of hodgkin - huxley neurons .
neuron plasticity is a flexible property of a neuron and its network to change temporarily or permanently their biochemical , physiological , and morphological characteristics , in order to adapt to the environment .
regarding the plasticity , we consider hebbian rules , specifically for spike timing - dependent plasticity ( stdp ) , and with regard to network , we consider that the connections are randomly distributed .
we analyse the synchronisation and desynchronisation according to an input level and probability of connections .
moreover , we verify that the transition for synchronisation depends on the neuronal network architecture , and the external perturbation level .
plasticity , neuronal network , synchronisation |
given @xmath2 , the riesz @xmath3energy of a set on points @xmath4 on a subset @xmath5 is @xmath6 this energy has a physical interpretation for some particular values of @xmath3 , i.e. for @xmath7 the riesz energy is the coulomb potential and for @xmath8 ( @xmath9 ) is the newtonian potential .
in the special case @xmath10 the energy is defined by @xmath11 and is related to the transfinite diameter and the capacity of the set by classical potential theory , see for example @xcite .
the minimal value of this energy and its asymptotic behavior have been extensively studied , most remarkably in the case that @xmath12 is the @xmath13dimensional unit sphere . in @xcite
it was proved that for @xmath14 and @xmath15 there exist constants @xmath16 ( depending only on @xmath13 and @xmath3 ) such that @xmath17 where @xmath18 is the continuous s - energy for the normalized lebesgue measure , @xmath19 finding the precise value of the constants in is an important open problem and has been addressed by several authors , see @xcite for some very precise conjectures and @xcite or @xcite for surveys .
one can post the problem as follows for @xmath20 , let @xmath21 be defined by @xmath22 find asymptotic values for @xmath21 as @xmath23 .
in particular , prove if the limit exists .
a sometimes successful strategy for the upper bound in the constant @xmath21 is to take collections of random points in @xmath24 and then compute the expected value of the energy ( which is of course greater than or equal to the minimum possible value ) .
simply taking @xmath25 points with the uniform distribution in @xmath24 already gives the correct term @xmath26 , and other distributions with nice separation properties have proved successful in bounding the constant @xmath21 .
we are thus interested in computationally feasible random procedures to generate points in sets which exhibit local repulsion .
one natural choice is using _ determinantal point processes _ which have these two properties ( see @xcite for theoretical properties and @xcite for an implementation ) .
a brief summary of the fundamental properties of determinantal point processes is given in section [ sec : dpp ] .
in a recent paper @xcite a determinantal point process named the _ spherical ensemble _ is used to produce low energy random configurations on @xmath27 .
this process was previously studied by krishnapur @xcite who proved a remarkable fact : the spherical ensemble is equivalent to taking eigenvalues of @xmath28 ( where @xmath29 have gaussian entries ) and sending them to the sphere through the stereographic projection . in @xcite
a different determinantal point process rooted on the use of spherical harmonics is described , producing low energy random configurations in @xmath24 for some infinite sequence of values of @xmath25 . in particular , it is proved in that paper that @xmath30 if @xmath31 , @xmath32 can be changed to @xmath33 in ( see ( * ? ? ? * cor .
the bound in is the best known to the date for general @xmath13 ( although more precise bounds exist for particular values of @xmath13 including @xmath34 , see @xcite ) . in particular , for @xmath35 and odd dimensions the formula in reads @xmath36 the determinantal point process in @xcite is called the _ harmonic ensemble _ and it is shown to be the optimal one ( at least for @xmath35 ) among a certain class of determinantal point processes obtained from subspaces of functions with real values defined in @xmath24 .
however , the bound in @xcite for the case @xmath37 is worse than that of @xcite , which is not surprising since the spherical ensemble uses complex functions and is thus of a different nature . an alternative natural interpretation of krishnapur s result is to consider eigenvalues @xmath38 of the generalized eigenvalue problem @xmath39 and to identify @xmath40 with the riemann sphere .
an homotety then generates the points in the unit sphere @xmath27 .
this remark suggests that the spherical ensemble can be seen as a natural point process in the complex projective space , and a search for an extension to higher dimensions is in order . in this paper
we extend this process in a very natural manner to @xmath41 for any @xmath42
. we will propose the name _ projective ensemble_. in order to show the separation properties of the projective ensemble we will define a ( probably non determinantal ) point process in odd dimensional spheres , which will allow us to compare our results to those of @xcite .
this point process is as follows : first , choose a number @xmath43 of points in @xmath41 coming from the projective ensemble .
then , consider @xmath44 equally spaced unit norm affine representatives of each of the projective points .
we allow these points to be rotated by a randomly chosen phase . as a result
, we get @xmath45 points in the odd dimensional sphere @xmath46 .
we study the expected @xmath1energy of such a point process .
our first main result can be succinctly written as follows .
[ th : main ] with the notations above , @xmath47 the bound in theorem [ th : main ] is larger than that of , which shows that random configurations of points coming from this point process are , at least from the point of view of the @xmath1energy , better distributed than those coming from the harmonic ensemble .
see figure [ fig : comparacion ] for a graphical comparison of both bounds .
since the point process we have defined in @xmath48 starts by choosing points in @xmath41 coming from the projective ensemble , theorem [ th : main ] gives us arguments to think that the projective ensemble produces quite well distributed points in @xmath41 ( for @xmath49 this property is quantitatively described in @xcite ) .
there are several ways to measure how well distributed a collection of points is in @xmath41 .
for example , one can study the natural analogues of riesz s energy as in theorem [ prop_proy ] below .
a very natural measure is given by the value of _ green s energy _ of @xcite : let @xmath50 $ ] be the green function of @xmath41 , that is , @xmath51 is zero
mean for all @xmath52 , @xmath53 is symmetric and @xmath54 , with @xmath55 the dirac s delta function , in the distributional sense .
the green energy of a collection of @xmath43 points @xmath56 is defined as @xmath57 minimizers of green s energy are assymptotically well distributed ( see ( * ? ? ?
* main theorem ) ) .
our second main result will follow from the computation of the expected value of green s energy for the projective ensemble .
[ th : main2 ] let @xmath58 . then , @xmath59 theorem [ th : main2 ] gives a criterium to decide how well distributed a collection of projective points is : compute their green s energy and compare to that of .
in this section we follow @xcite .
[ def : pp ] let @xmath60 be a locally compact , polish topological space with a radon measure @xmath61 . a _ simple point process _
@xmath62 of @xmath25 points in @xmath60 is a random variable taking values in the space of @xmath25 point subsets of @xmath60 .
there are some subtle issues in the general definition of point processes , see ( * ? ? ?
* section 1.2 ) . for our purposes
we will only use simple point processes with a fixed , finite number of points .
for some point processes there exist _ joint intensities _ satisfying the following definition .
let @xmath63 be as in definition [ def : pp ] .
the joint intensities are functions ( if any exist ) @xmath64 , @xmath65 such that for any family of mutually disjoint subsets @xmath66 of @xmath60 we have @xmath67 here , @xmath68 denotes expectation and by @xmath69 we mean that @xmath52 is a subset of @xmath60 with @xmath25 elements , obtained from the point process @xmath62 . from ( * ? ? ?
* formula ( 1.2.2 ) ) , for any measurable function @xmath70 the following equality holds .
@xmath71 sometimes these intensity joint functions can be written as @xmath72 for some function @xmath73 . in this case
, we say that @xmath62 is a _
determinantal point process_. a particularly amenable collection of such processes is obtained from @xmath25dimensional subspaces of the hilbert space @xmath74 ( i.e. the set of square
integrable complex functions in @xmath60 ) .
recall that the reproducing kernel of @xmath75 is the unique continuous , skew symmetric , positive definite function @xmath76 such that @xmath77 given any orthonormal basis @xmath78 of @xmath75 , we have @xmath79 such a kernel @xmath80 is usually called a _ projection kernel _ of trace @xmath25 .
[ prop : ms ] let @xmath60 be as in definition [ def : pp ] and let @xmath81 have dimension @xmath25 .
then there exists a point process @xmath82 in @xmath60 of @xmath25 points with associated join intensity functions @xmath83 in particular for any measurable function @xmath84 we have @xmath85 we will call @xmath82 a projection determinantal point process with kernel @xmath80 . this proposition is a direct consequence of the macchi soshnikov theorem ,
see @xcite or ( * ? ? ?
* theorem 4.5.5 ) .
@xmath86 [ obs : constant ] we note that in the hypotheses of proposition [ prop : ms ] , from with @xmath87 we also have @xmath88 in particular , if @xmath89 is constant then we must have @xmath90 .
we now describe the push forward of a projection determinantal point process .
we are most interested in the case that the spaces are riemannian manifolds ( which are locally compact , polish and measurable spaces ) .
[ prop_1 ] let @xmath91 and @xmath92 be two riemannian manifolds and let @xmath93 be a @xmath94 diffeomorphism .
let @xmath95 be an @xmath25dimensional subspace .
then , the set @xmath96 is an @xmath25dimensional subspace of @xmath97 .
its associated determinantal point process @xmath98 has kernel @xmath99 ( we are denoting by @xmath100 the jacobian determinant ) .
this proposition is a direct consequence of the change of variables formula , see section [ subsec : proof1 ] for a short proof .
consider the standard fubini study metric in the complex projective space of complex dimension @xmath13 , denoted by @xmath41 .
the distance between two points @xmath101 is given by : @xmath102 [ def_2 ] let @xmath103 and consider the set of the following functions defined in @xmath104 : @xmath105 where @xmath106 are non negative integers and @xmath107 let @xmath108 which is a subspace of complex dimension @xmath109 .
the collection @xmath110 given in is an orthonormal basis of @xmath111 ( for the usual lebesgue measure in @xmath104 ) and the reproducing kernel @xmath112 is given by : @xmath113 from proposition [ prop : ms ] , there is an associated determinantal point process of @xmath43 points in @xmath114 that we denote by @xmath115 .
[ propdef_1 ] let @xmath116 and let @xmath43 be of the form @xmath117 .
then , the pushforward @xmath118 of @xmath115 under the mapping @xmath119 is a determinantal point process in @xmath41 whose associated kernel satisfies @xmath120 we call this process the projective ensemble
. see section [ proplemma32 ] for a proof of lemma [ propdef_1 ] .
the spherical ensemble described in @xcite is just the case @xmath49 of the projective ensemble identifying @xmath121 with the riemmann sphere and translating the process to the unit sphere .
the next result computes the expected value of a riesz like energy for the projective ensemble .
[ prop_proy ] let @xmath122 . for @xmath123 and @xmath124
let @xmath125 then , for @xmath126 , @xmath127 note that @xmath128 is precisely the continuous @xmath3energy for the uniform measure in @xmath41 .
[ cor : logaritmica ] let @xmath122 .
for @xmath123 and @xmath124 let @xmath129 then , @xmath130 theorem [ prop_proy ] and corollary [ cor : logaritmica ] are proved in sections [ sec : proofofprop_proy ] and [ sec : prooflog ] .
we now describe a point process of @xmath25 points , for certain values of @xmath25 , in @xmath131 in the following manner .
[ def : pps2dmas1 ] given integers @xmath132 , let @xmath123 and @xmath133 .
we define the following point process of @xmath25 points in @xmath131 .
first , let @xmath134 be chosen from the projective ensemble @xmath118 .
choose , for each @xmath135 , one affine representative ( which we denote by the same letter ) .
then , let @xmath136 be chosen uniformly and independently and define @xmath137 we denote this point process by @xmath138 . note that the way to generate a collection of @xmath25 points coming from @xmath138 amounts to taking @xmath43 points from the projective ensemble and taking , for each of these points , @xmath44 affine unit norm representatives , uniformly spaced in the great circle corresponding to each point , with a random phase .
the following statement shows that the expected @xmath1energy of points generated from the point process of definition [ def : pps2dmas1 ] can be computed with high precision .
it will be proved in section [ sec : proofuju1 ] .
[ uju1 ] @xmath139 following the same ideas one can also compute the expected @xmath3-energy for @xmath25 points coming from the point process @xmath138 for other even integer values @xmath140 , and a bound can be found for other values of @xmath141 .
the computations , though , are quite involved .
proposition [ uju1 ] describes how different choices of @xmath142 ( i.e. of @xmath43 ) and @xmath44 produce different values of the expected @xmath1energy of the associated @xmath143 points .
an optimization argument is in order : for given @xmath25 , which is the optimal choice of @xmath43 and @xmath44 ? since we know from that the second order term in the assymptotics is @xmath144 , it is easy to conclude that the optimal values of @xmath43 and @xmath44 satisfy : @xmath145 the following corollary follows inmediately from proposition [ uju1 ] .
[ cor : casi ] if we choose @xmath146 for some @xmath147 making that quantity a positive integer , then : @xmath148 the proof of our first main theorem will follow easily from corollary [ cor : casi ] .
we first prove that @xmath149 . indeed ,
for @xmath150 we have @xmath151 for some @xmath152 . since it is in one to
one correspondence with @xmath75 , the dimension of @xmath153 is also @xmath25 . now , by the change of variables formula this last equals the squared @xmath154 norm of @xmath155 which is finite since @xmath156 .
we now prove the formula for @xmath157 .
let @xmath158 be an orthonormal basis of @xmath75 .
then , @xmath159 , @xmath160 , are elements in @xmath153 and using the change of variables formula we have : @xmath161 where we use the kronecker delta notation .
hence , @xmath162 form an orthonormal basis and @xmath163 the other formula for @xmath164 follows from this last one , using that @xmath165 let @xmath172 be the quantity we want to compute . following proposition [ prop : ms ] we have that @xmath173 where we choose unit norm representatives @xmath174 . since the integrand only depends on the distance between @xmath175 and @xmath176 and @xmath41
is a homogeneous space , we can fix @xmath177 to get : @xmath178 where we have used that the volume of @xmath41 is equal to @xmath179 . in order to compute this integral
, we use the change of variables theorem with the map @xmath167 whose jacobian is given in , getting : @xmath180 integrating in polar coordinates , @xmath181 \\ & = 2r^{2}d \left [ \frac{b \left ( d - \frac{s}{2 } , 1 \right)}{2 } - \frac{b \left ( d - \frac{s}{2 } , l + 1 \right)}{2 } \right ] \\ & = \frac{d}{d - \frac{s}{2}}r^{2 } - r^{2}db \left ( d - \frac{s}{2 } , l + 1 \right ) , \end{split}\ ] ] note that @xmath186 . in particular , interchanging the order of expected value and derivative ( it is an exercise to check that this change is justified ) , from theorem [ prop_proy ] we have @xmath187 the proof of the corollary is now a straightforward computation of that derivative and it is left to the reader .
it is helpful to recall the derivative of euler s beta function in terms of the digamma function @xmath188 for @xmath189 : @xmath190 @xmath193 } \esp{\xr}{\e{2}(y_{1}^{0}, ... ,y_{1}^{k-1},y_{2}^{0}, ... ,y_{r}^{k-1 } ) } d(\theta_{1}, ... ,\theta_{r } ) \\ & = \frac{1}{(2\pi)^{r } } \int_{\theta_{1}, ...
,\theta_{r } \in [ 0 , 2\pi ] } \esp{\xr } { \sum_{i_{1 } \neq i_{2}\text { or } j_{1 } \neq j_{2 } } \frac{1}{\left|\left| y_{i_{1}}^{j_{1 } } - y_{i_{2}}^{j_{2 } } \right|\right|^ { 2 } } } d(\theta_{1}, ... ,\theta_{r } ) \\ & = j_1+j_2 , \end{split}\ ] ] where @xmath194 } \esp{\xr } { \sum_{i=1}^{r } \sum_{j_{1 } \neq j_{2 } } \frac{1}{\left|\left| y_{i}^{j_{1 } } - y_{i}^{j_{2 } } \right|\right|^ { 2}}}\,d(\theta_{1}, ... ,\theta_{r } ) , \\ j_2= & \frac{1}{(2\pi)^{r } } \int_{\theta_{1}, ...
,\theta_{r } \in [ 0 , 2\pi]}\esp{\xr } { \sum_{j_{1},j_2=0}^{k-1 } \sum_{i_{1 } \neq i_{2 } } \frac{1}{\left|\left| y_{i_{1}}^{j_{1 } } - y_{i_{2}}^{j_{2 } } \right|\right|^ { 2 } } } \ , d(\theta_{1}, ... ,\theta_{r}).\end{aligned}\ ] ] from we have : @xmath195 } \esp{\xr } { \sum_{j_{1 } \neq j_{2 } } \frac{1}{\left|\left| e^{\i\left ( \theta + \frac{2\pi j_1}{k } \right ) } x_i - e^{\i\left ( \theta + \frac{2\pi j_2}{k } \right ) } x_i \right|\right|^ { 2}}}\,d\theta.\end{aligned}\ ] ] now , the integral does not depend on @xmath196 nor in the ( unit norm ) vector @xmath197 , so we actually have that @xmath198 is the @xmath1energy of the @xmath44 roots of unity .
this quantity has been studied with much more detail than we need in ( * ? ? ?
* theorem 1.1 ) .
in particular , we know that it is of the form @xmath199 . we thus conclude : @xmath200 we now compute @xmath201 .
interchanging the order of integration we have : where we can choose whatever unit norm representatives we wish of @xmath203 and @xmath204 . in order to compute the inner integral , for any fixed @xmath205 we assume that our choice satisfies @xmath206 $ ] ( i.e. it is real and non negative ) , which readily implies @xmath207 a simple computation using the invariance of the integral under rotations yields : @xmath208 and this last value is independent of @xmath209 .
we thus have : @xmath210 this last expected value has been computed in theorem [ prop_proy ] , which yields : @xmath211 proposition [ uju1 ] follows from , and .
fix @xmath116 and let @xmath212 be the coefficient of @xmath213 in .
the function @xmath214 has a strict global minimum at @xmath215 indeed , @xmath216 gives the bound for the @xmath32 given in theorem [ th : main ] .
we can not just let @xmath217 in corollary [ cor : casi ] since it might happen that @xmath218 , but we will easily go over this problem .
let @xmath122 be any positive integer , let @xmath123 and let @xmath219 be the unique number in the interval @xmath220 such that @xmath221 .
finally , let @xmath222 , which depends uniquely on @xmath13 and @xmath142 , and which satisfies @xmath223 as @xmath182 .
for any @xmath224 we then have : @xmath225 the first inequality from corollary [ cor : casi ] and the second inequality due to @xmath183 as @xmath182 , which implies for some constant @xmath226 : @xmath227 we have thus proved @xmath228 which finishes the proof of our theorem [ th : main ] . from @xcite , the green function of @xmath41 is given @xmath229 where @xmath230 and @xmath231 integrating the formula above ( see for example @xcite we have : @xmath232 + c. \end{split}\ ] ] in order to compute the constant we need to impose that the average of @xmath51 equals @xmath233 for all ( i.e. for some ) @xmath234 .
let @xmath235 and change variables using @xmath167 from lemma [ propdef_1 ] whose jacobian is given in to compute : @xmath236.\ ] ] integrating in polar coordinates , @xmath237 ( for the computation of the integrals , use the change of variables @xmath238 and @xcite , for example ) .
@xmath241 + o\left(r^{2-\frac{1}{d}}\right)\\ & = r^{2 } \frac{d!}{4\pi^{d } } \left ( \sum_{k=1}^{d-1 } \frac{1}{k(d - k ) } + \frac{1}{d^2 } \right ) - \frac{(d!)^{1- \frac{1}{d}}}{4\pi^{d}(d-1 ) } r^{2 - \frac{1}{d } } + o\left(r^{2-\frac{1}{d}}\right ) \\ & = r^{2 } \frac{(d-1)!}{4\pi^{d } } \left ( \frac{1}{d}+2\sum_{k=1}^{d-1 } \frac{1}{k } \right ) - \frac{(d!)^{1- \frac{1}{d}}}{4\pi^{d}(d-1 ) } r^{2 - \frac{1}{d } } + o\left(r^{2-\frac{1}{d}}\right ) \end{split}\ ] ] | we define a determinantal point process on the complex projective space that reduces to the so called spherical ensemble for complex dimension @xmath0 under identification of the @xmath1sphere with the riemann sphere . through this determinantal point process
we propose a point processs in odd - dimensional spheres that produces fairly well distributed points , in the sense that the expected value of the riesz @xmath1energy for these collections of points is smaller than all previously known bounds . |
mixing is one of the most important process in industries , including polymer processing , rubber compounding , and food processing , because it directly affects the quality of multi - component materials @xcite .
several types of mixing devices , such as twin - screw extruders , twin - rotor mixers , and single - screw extruders , have been developed for different material processabilities and different processing purposes . to select an appropriate mixing element for a variety of complex materials and product qualities ,
one fundamental issue is the fluid mechanical characterization of the mixing capability of various types of the mixing elements . the mixing process , or ,
more specifically , the reduction of the inhomogeneity of material mixtures , is achieved through material flow driven by mixing devices .
therefore , in principle , the capability of the mixing elements can be evaluated through an analysis of the flow in the device . along these lines ,
the visualization of the flow patterns has been performed numerically or experimentally to obtain a qualitative insight into the global mixing kinetics @xcite . while the global flow pattern characterizes the evolution of the material distribution , the divergence of the material trajectories
is locally associated with the deformation of the fluid elements . since the substantial local deformation rate is described by the strain rate tensor ,
several approaches for the characterization of the local strain rate have been developed @xcite .
the degree of irrotationality of the deformation rate , which is often called the `` mixing index '' @xcite , has been used to quantify dispersive mixing efficiency @xcite .
this characterization was motivated by the experimental fact that elongational flows are more effective than simple shear for droplet / agglomerate breakup . in a rheometric flow setup ,
the elongational flows are irrotational , while the simple shear flow is a superposition of a planar shear flow and a rotational flow in a plane .
therefore , the irrotationality can define the flow type for the rheometric flow setup .
however , in generic flows in mixing devices , the type of elongational flows described by the strain rate tensor can not be assessed by the irrotationality by definition .
another quantity , called the mixing efficiency , is the relative magnitude of the elongational rate along a certain direction to the magnitude of the strain rate tensor @xcite .
the mixing efficiency is useful when the interface between phases is well defined .
if , in the mixing efficiency , the maximal elongational direction is taken , the modes of the strain rate , such as the uniaxial / biaxial elongational flows and the planar shear flow , are discriminated @xcite . in this case , the eigenvalue problem for the strain rate tensor should be solved .
the distribution of the strain - rate modes in combination of the magnitude of the strain - rate in general flows should be useful in understanding the dispersive mixing capability .
the strain - rate modes in three - dimensional flows can be identified in principle by combining the different eigenvalues of the strain - rate tensor , and thus different forms for quantification of the strain - rate mode can be designed .
however , such quantification has not been derived in a numerically tractable manner . to the best of our knowledge
, the role of elongational flow in distributive mixing has been only rarely discussed , whereas the effectiveness of irrotational elongational flows for dispersive mixing has been well recognized .
flow trajectories diverge by elongational flows however small the elongational rate is .
the divergence directions are restricted to a plane for planar shear flow , but involve three directions for non - planar elongational flows .
if distributive mixing is effectively promoted in the confined space of a mixing device during a finite period of time , a flow pattern being effective to distributive mixing should be developed .
such a flow pattern is expected to be associated with a certain distribution of the non - planar or volumetric elongational flows .
the distribution of the volumetric flows is therefore considered to give essential information for a better understanding of the mixing capability of mixing elements .
especially , the flow patterns in the regions of smaller strain rates largely determine net mixing capabilities of mixing devices .
although the area - stretching ability is very low in the small strain rate regions , such regions occupy a large fraction of the channel , and the materials are conveyed to larger strain rate regions by the flow in the small strain rate regions .
the distribution of the volumetric flows is expected to be useful in understanding the relation between the structure of the flow patterns in the small strain regions and the mixer geometry .
in general , the flow field in a mixing device is an arbitrary combination of volumetric elongational flows and a planar shear flow .
thus , for a general three - dimensional flow , rather than only rheometric flows , the characterization of volumetric elongational flows from the strain rate tensor is required .
although the strain - rate mode should be an important quantity both in dispersion capability and analysis of flow - pattern structures in relation to distributive mixing , its characterization in three - dimensional flows has not established so far . in this paper , we derive a measure which identifies the volumetric elongational flows from the strain rate tensor .
we apply this measure to melt - mixing flow in twin - screw extrusion . using the distribution of the volumetric elongational flows , we discuss the differences in the flow patterns and the mixing characteristics of the different screw elements .
in any kind of flow in a mixing device , the local deformation rate is a combination of shear flow , elongational flow , and rotational flow .
the deformation rate of a fluid element , @xmath0 , is decomposed into the strain rate tensor @xmath1 and the vorticity tensor @xmath2 : @xmath3 where @xmath4 is the velocity field , @xmath5 , and @xmath6 , the superscript @xmath7 indicates the transpose .
concerning mixing processes , the change in the distance between two nearby points is represented by @xmath1 .
thus , the strain rate tensor is mainly responsible for the local mixing capability . for incompressible flows , _
i.e. _ , @xmath8 , the strain rate tensor can be diagonalized with an orthonormal matrix @xmath9 , @xmath10 where @xmath11 and @xmath12 , respectively , represent the largest and the smallest eigenvalues of @xmath1 by assuming @xmath13 .
the value of @xmath14 specifies the mode of the strain rate .
for example , @xmath15 for a uniaxial elongational flow , @xmath16 for a biaxial elongational flow , and @xmath17 for a planar shear flow . a typical value of the strain rate is @xmath18 , but the magnitude of the strain rate is commonly evaluated by the second invariant of @xmath1 without explicit calculation of the eigenvalues , @xmath19 because it is always positive for finite values of @xmath18 by definition .
since we are interested in the mode of the strain rate , or , equivalently , the value of @xmath14 without solving the eigenvalue problem , we consider another invariant of @xmath1 .
the determinant of @xmath1 , a third - order invariant , is expressed in terms of the eigenvalues by @xmath20 from eq .
( [ eq : determinant ] ) , we can see that the determinant of @xmath1 becomes zero for planar shear ( @xmath17 ) , irrespective of @xmath18 .
in other words , @xmath21 means that the strain directions are restricted to be within a certain plane , and @xmath22 indicates that the strain directions extend three - dimensionally . with this property of @xmath23
, we can get an insight into whether the local strain rate is more planar or more volumetric without explicitly calculating the eigenvalues of @xmath1 . combining eqs .
( [ eq : dd ] ) and ( [ eq : determinant ] ) , we define a measure for the mode of the strain rate , independent of the value of @xmath18 , by @xmath24 where the numerical prefactor is chosen so that the range of @xmath25 is normalized to @xmath26 $ ] .
special cases include : @xmath27 for a biaxial elongational flow , @xmath28 for a uniaxial elongational flow , and @xmath29 for a planar shear flow .
for the general case of @xmath30 , the strain rate is an arbitrary superposition of an elongational flow and a shear flow .
both @xmath31 and @xmath23 are frame invariant , implying that @xmath25 is frame invariant as well , namely @xmath25 does not change in different frames of the observers .
now we consider a relation between the strain rate mode and the mixing capability . in general ,
if a mixing device has good mixing capability , the flow in the device should satisfy the following two conditions .
( i ) there exist some regions with large strain rates in order for the materials to be stretched extensively .
( ii ) all the fluid elements in the mixer are efficiently and repeatedly conveyed to the large strain rate regions in order to have chances to be stretched .
the distribution of the magnitude of the strain rate of eq .
( [ eq : dd ] ) ( or in the form of shear stress , @xmath32 , with the shear viscosity @xmath33 as a function of the magnitude of the strain rate ) has often been discussed @xcite . in mixing devices for highly viscous fluids , like polymers ,
the largest strain rate is achieved in narrow gap regions , which are the clearance between the screw / rotor tips and the barrel surface , and the gap between two screws / rotors in twin - shaft machines .
thus , it is usually the case the condition ( i ) is satisfied .
however , it is only a necessary condition for good net mixing capability .
these large strain rate regions are also responsible for the dispersion of agglomerates .
in contrast to the large strain rate regions , a role of regions with smaller strain rates on net distributive mixing is rather difficult to characterize because this would , in principle , require an analysis of the ensemble of trajectories . nevertheless , the small strain rate regions are mainly responsible for the condition ( ii ) . in the past ,
distributive mixing was discussed using tracer tracking and lagrangian statistics @xcite .
these methods are basically for the characterization of the global properties of a mixing device , and are not for the direct characterization of a relation between the local flow characteristics induced by a screw geometry and the mixing capability .
effective and repetitive transport of the materials to the large strain rate regions is one important role of the small strain rate region , and it affects net mixing performance of the mixers . this flow pattern structure in the small strain rate regions is basically based on the mixer geometry . in order for the fluid elements to be stretched and folded over the whole space in a mixing device , the flow trajectories should be bifurcated and change directions with some frequency during the mixing process .
this bifurcation of the flow trajectories is directly related to the volumetric modes of the strain rate .
therefore , the spatial distribution of the strain rate mode gives directly an insight into the relation between the mixing process and the flow driven by the mixer geometry .
for the characterization of the local deformation rate , there is another quantity , called the mixing index @xcite , @xmath34 which has often been used to characterize the efficiency of the dispersive mixing of mixing devices @xcite .
some authors have used different normalization like @xmath35 which takes a value in @xmath26 $ ] @xcite while @xmath36 is normalized to be within @xmath37 $ ] .
since @xmath38 is not frame invariant when the observer s frame depends on time , @xmath36 is not frame invariant in general .
nevertheless , if we choose a time independent frame like the laboratory system of coordinates , @xmath36 is practically regarded as objective . by the definition in eq .
( [ eq : mz ] ) , @xmath36 defines the degree of irrotationality of the deformation rate , _
e.g. _ , @xmath39 for pure rotation , @xmath40 for an irrotational flow , and @xmath41 for a partially rotational flow ; furthermore , @xmath36 is independent of the strain rate mode . in two - dimensional flows , since available strain - rate mode is only planar elongational flow , irrotationality can be used to distinguish between simple shear flow ( @xmath42 ) and elongational flow ( @xmath40 ) . however , this is not the case for three - dimensional flows since different strain - rate modes exist . in three - dimensional flows ,
@xmath36 simply indicates the irrotationality , but not the strain - rate mode .
the mixing index @xmath36 and our measure for the strain rate mode @xmath25 characterize different aspects of the local deformation rate .
the flow of a polymer melt in the melt - mixing zone of a twin - screw extruder has been numerically solved to allow understanding the relation between the distribution of the strain rate mode and the mixing capability of screw elements .
the configuration of the screws is shown in fig .
[ fig1 ] . from the inlet ,
a forwarding conveying screw , kneading disks , and a backwarding conveying screw are arranged .
the melt - mixing zone consists of kneading disks of neutrally staggered five blocks @xcite .
the diameter of the barrel is set to @xmath43 mm , the length of the computational domain is @xmath44 .
we focus on the situation where the material fully fills the channel .
the reynolds number is assumed to be much less than unity , so that inertial effects are neglected .
the flow is assumed to be incompressible , and in a pseudo - steady state to screw rotation , as has often been assumed in polymer flow in twin - screw extruders @xcite . with these assumptions ,
the governing equations become @xmath45 where @xmath46 is the pressure , @xmath47 is the deviatoric stress , @xmath48 is the mass density , @xmath49 is the specific heat capacity , @xmath7 is the temperature , and @xmath50 is the thermal conductivity .
the fluid is assumed to be a viscous shear - thinning fluid that follows cross arrhenius viscosity @xcite , @xmath51 whose parameters are obtained by fitting the shear viscosity of a polypropylene melt taken from @xcite , and the values are @xmath52 , @xmath53 , @xmath54 , and @xmath55 . the mass density , specific heat capacity , and thermal conductivity are taken from @xcite as well , and the values are @xmath56kg / m@xmath57 , @xmath58j/(kg@xmath59k ) , and @xmath60w/(m@xmath59k ) . as operational conditions , the volume flow rate and the screw rotation speed are set to 60cm@xmath57/s ( @xmath62kg / h ) and 200rpm .
the no - slip condition on the velocity at the barrel and screw surfaces is assumed .
the velocity at the inlet boundary was set to a value under the given volume flow rate .
the pressure at the outlet boundary was fixed to be a constant value .
the temperatures on the barrel surface and at the inlet boundary were set to be a constant value of 473.15k .
the natural boundary conditions for the temperature equation in the exit boundary plane and the screw surface were assigned .
the set of equations was discretized by the finite volume method and solved by simple method @xcite using a commercial software , `` r - flow '' ( r - flow co. , ltd . ,
saitama , japan ) . from the obtained velocity field ,
the strain rate and the vorticity tensors were calculated so as to evaluate @xmath25 and @xmath36 .
for typical twin - screw mixers , the largest strain rate occurs mainly around the tips and the gap between the two screws in the intermeshing region because the channel widths of these regions are designed to be the smallest in the device .
this is common for the conveying screw of extruders and many types of kneading elements , such as kneading disks , rotor elements , screw mixing elements , and so on @xcite . to observe this , fig .
[ fig2 ] shows the distribution of the magnitude of the strain rate , @xmath63 , at the cross section b indicated in fig .
[ fig1 ] . at the clearances between the screw - tips and the barrel surface , and the gap between two screws
, @xmath63 takes values larger than 600@xmath64/s ) , which is consistent with the circumferential shear rate of 460@xmath64/s ) estimated from the screw rotation speed of 200rpm .
in contrast , the magnitude of the strain rate in other regions is extremely small .
the distribution of @xmath63 is basically similar for the conveying screws and the kneading disks shown in fig .
[ fig1 ] , suggesting that the difference in the mixing capabilities of these elements are largely attributable to the differences in the flow patterns in the small strain rate regions . .
large values of \protect\sqrt{\protect{\mbox{\boldmath $ d$}}:\protect{\mbox{\boldmath $ d$}}}\protect over 600(1/s ) are reached at the narrow gaps between the screw tips and the barrel surface and between the two screws , while in the other regions , the value of \protect\sqrt{\protect{\mbox{\boldmath $ d$}}:\protect{\mbox{\boldmath $ d$}}}\protect stays rather small . ` ] next , we compare the distributions of the volumetric strain rate between the conveying screw and the kneading disks .
we use the absolute value of @xmath25 as defining @xmath65 , which is zero for a planar shear , unity for pure elongational flows , and becomes @xmath66 for an arbitrary superposition of planar and elongational flows , for we are interested in how the volumetric strain rate is distributed .
figure [ fig3 ] shows the distributions of @xmath67 at a cross section in the conveying screw ( location b in fig .
[ fig1 ] ) and at a cross section in the kneading disks ( location a in fig .
[ fig1 ] ) . in the case of the conveying screw shown in fig .
[ fig3](a ) , although an elongation flow occurs in some small regions around the tip - barrel clearance and in the intermeshing regions , small values of @xmath67 prevail in most of the section .
this was observed in the other phase of the screw rotation . from this observation ,
circumferential planar shear is predominant in the flow driven by a conveying screw , which corresponds to the flow along the screw root .
the elongational flows around the tip - clearance regions in the conveying screw are explained as follows .
when fluid elements go across a screw tip , that flow is a bifurcation from the upper stream flow along the screw root , followed by the confluence to another flow along the next screw root .
these flow patterns are observed as the elongational flows in fig .
[ fig3](a ) . except for these flow patterns ,
interchange of materials rarely occurs along screw rotations .
this fact is consistent with the well - known low level of the mixing capability of a conveying screw , because a volumetric bifurcation of the trajectories rarely occurs , so that the distributive mixing is not much promoted . in the case of the kneading disks shown in fig .
[ fig3](b ) , we found that the volumetric strain rate develops in a remarkably large fraction of the section , and its distribution forms a characteristic pattern .
elongational flows occur in the region far from the screw tips and the surfaces , which are , coincidentally , small strain rate regions .
the locations of the elongational flows correspond to those of the opening space between neighboring staggered disks . as demonstrated in fig .
[ fig3 ] , the distribution of the volumetric strain rate closely reflects the flow pattern structure caused by the geometric shapes of the elements , and characterizes the flow pattern in the small strain rate regions . , ( b ) the section in the kneading disks at position a in fig .
[ fig1 ] . ` ] as for the kneading disks , the characteristic pattern in the distribution of @xmath67 is supposed to be related to their known good mixing capability @xcite . for understanding the relation between the distribution of @xmath67 and the mixing capability
, we discuss the flow pattern in the small strain rate region in the kneading zone .
figure [ fig4](a ) shows the velocity field at the mid - plane of the channel at the third disk in the kneading zone . from fig .
[ fig4](a ) , we see the flow trajectories from a tip to two neighboring disks along the screw rotation . simultaneously , converging flows from two neighboring disks
are developed behind another side of screw tips .
in other words , the flow driven by the kneading disks bifurcates into forward and backward extrusion directions on the one place and converges from neighboring disks on the other place .
because of these bifurcated trajectories , the fluid elements go back and forth within the zone of consecutively arranged disks , and repeatedly bifurcate and converge having many chances to be stretched and folded resulting in the high mixing capability of the kneading disks .
this property of the flow pattern is reflected by the distribution of the biaxial / uniaxial elongational flow from the viewpoint of the strain rate mode .
figure [ fig4](b ) shows the distribution of @xmath67 corresponding to the velocity field in fig .
[ fig4](a ) and clearly captures this characteristic flow pattern structure developed in the small strain rate region of the kneading disks .
this analysis demonstrates that the volumetric strain rate distribution can be useful to characterize the flow pattern structure , especially in the small strain rate regions . for comparison , the distribution of the mixing index ( or the irrotationality ) , @xmath36 , of eq .
( [ eq : mz ] ) at the cross sections a and b in fig
. [ fig1 ] are shown in fig .
we found that both the conveying screw and the kneading disks share some common characteristics in the @xmath36 distribution .
the deformation rate is almost half rotational near the surfaces of the screws and the barrel , while it is almost irrotational in the small strain rate region far from the surfaces .
although the irrotational regions for both elements are different , depending on their geometric shapes , the @xmath36 distribution does not reflect the flow pattern shown in fig .
[ fig4](a ) , suggesting that the mixing index itself hardly offers insight into the flow pattern structure in general three - dimensional flow . ,
( b ) the section in the kneading disks at position a in fig .
[ fig1 ] . ` ] in order to discuss the large scale characteristics of the screw elements , @xmath67 and @xmath36 are averaged in each section and over one screw rotation .
the axial profiles of the mean @xmath67 and @xmath36 are shown in fig .
[ fig6 ] . in proximity to the inlet and the outlet , the mean
values are supposed to be affected by the boundary conditions , and so , physically irrelevant .
we hence consider the axial locations from 20 mm to 200 mm .
the means of @xmath67 and @xmath36 remain at an almost constant level in the regions of the two conveying screws .
in addition , these mean quantities do not vary much at the first and last blocks in the kneading disks because they are rather smoothly connected to the conveying screws .
in contrast , the means of @xmath67 and @xmath36 show large variation along the inner three blocks of the kneading disks . the piecewise variations of these values are due to the piecewise structure of the kneading disks .
in particular , the variation in the mean @xmath67 within the inner three blocks is remarkable .
the mean @xmath67 has a peak at the center of each of the three blocks , which reflects the bifurcated flow trajectories back and forth as observed in fig .
the axial profile of the mean @xmath67 clearly shows that the flow pattern structure specific to the kneading disks occurs in the inner blocks and originates from the consecutive staggered arrangement of the blocks .
as demonstrated above , a flow pattern being effective for distributive mixing is closely related to the distribution of the volumetric strain rate which indicates three - dimensional bifurcation and converging of trajectories .
the distribution of @xmath67 offers a physical insight into the high mixing capability of the kneading disks as well as the low mixing capability of the conveying screws .
we derived a scalar measure , @xmath25 , which characterizes the mode of the local strain rate by combining the second and the third invariants of the strain rate tensor . using the value of @xmath25
, the mode of the strain rate tensor including a uniaxial / biaxial elongational flow , a planar shear flow , and arbitrary combinations , is defined for the general three - dimensional flows observed in fluid mechanical processes without solving the eigenvalue problem for the strain rate tensor . the spatial distribution of non - planar , or volumetric , strain rate is closely related to the flow patterns , irrespective of the magnitude of the strain rate , and therefore it was found to be useful to understand the relation between the mixer geometry and the flow pattern structures . from the viewpoint of mixing processes , the flow pattern in the small strain rate regions has an important role in efficient and repetitive transport of the fluid elements to the large strain rate regions in order to enhance the net mixing capability of the mixer . for understanding the effectiveness of the flow pattern especially in the small strain rate regions , the distribution of the volumetric strain rates can be a useful tool .
based on the numerical simulation of a melt - mixing flow in twin - screw extrusion , flows driven by the conveying screws and the kneading disks have been analyzed .
we found that the flow patterns specific to these elements are clearly characterized by the distribution of the strain rate mode .
understanding the relation between the geometric structure of the mixing elements and the flow pattern they drive is an important issue in the essential evaluation of the mixing capability of the mixing devices used in different industries .
the analysis employing the strain rate mode and its distribution is effective for discussing the flow pattern in the mixing device and offers an insight into the role of the small strain rate regions on the distributive mixing .
finally , we would like to mention a limitation of the analysis only by @xmath25 for the mixing .
although the distribution of the volumetric strain rates can give information about the flow pattern structures specifying the regions where the flow bifurcates and converges , it can just distinguish the potential of particular flows . for direct evaluation of the mixing , kinetic evolution of interface
is needed to discuss the area growth , interface folding , and material distribution .
computation of the interface kinetics requires the directions of the area segments , and therefore the mixing is a function of the orientation of the interface relative to the flow .
this aspect of the mixing is another important problem from the evaluation of the potential of the flow . the combined use of our measure for the strain rate mode with other fluid mechanical analyses , including kinetic evolution of area elements is an important future direction of research into predicting the mixing capabilities of different mixing elements and in designing improved novel mixing elements .
the numerical calculations have been partly carried out using the computer facilities at the research institute for information technology at kyushu university . this work has been supported by grants - in - aid for scientific research ( jsps kakenhi ) under grants nos . 26400433 , 24656473 , and 15h04175 .
funatsu k , kihara si , miyazaki m , katsuki s , kajiwara t. numerical analysis on the mixing performance for assemblies with filled zone of right - handed and left - handed double - flighted screws and kneading blocks in twin - screw extruders .
_ polym eng sci_. 2002;42(4):707723 .
bravo vl , hrymak an , wright jd .
study of particle trajectories , residence times and flow behavior in kneading discs of intermeshing co - rotating twin - screw extruders .
_ polym eng sci_. 2004;44(4):779793 .
lawal a , kalyon dm .
simulation of intensity of segregation distributions using three - dimensional fem analysis : application to corotating twin screw extrusion processing .
_ j appl polym sci_. 1995;58(9):15011507 . | understanding the mixing capability of mixing devices based on their geometric shape is an important issue both for predicting mixing processes and for designing new mixers . the flow patterns in mixers
are directly connected with the modes of the local strain rate , which is generally a combination of elongational flow and planar shear flow .
we develop a measure to characterize the modes of the strain rate for general flow occurring in mixers .
the spatial distribution of the volumetric strain rate ( or non - planar strain rate ) in connection with the flow pattern plays an essential role in understanding distributive mixing . with our measure , flows with different types of screw elements in a twin - screw extruder are numerically analyzed .
the difference in flow pattern structure between conveying screws and kneading disks is successfully characterized by the distribution of the volumetric strain rate .
the results suggest that the distribution of the strain rate mode offers an essential and convenient way for characterization of the relation between flow pattern structure and the mixer geometry . |
fluids dissipate energy as they flow through pipes or past any smooth or rough surface .
examples include river flow or wind blowing across the land .
this energy dissipation is proportional to the velocity gradient , or shear rate , of the flow at the bounding surface .
this frictional energy loss , and its dependence on the reynolds number of the flow @xcite , is not yet fully understood a century after the first explanation was advanced by l. prandtl @xcite .
here we introduce a new scheme for measuring the shear rate near a bounding surface .
it also might be applicable in the interior of a fluid . unlike some widely - used methods @xcite , the shear rate @xmath3 is recorded at a single `` point '' of size @xmath4 .
the motivation for developing this technique was to improve the usual method for measuring the shear rate in turbulent flows @xcite .
the scheme introduced here is that of photon correlation spectroscopy ( pcs ) @xcite .
it is a variant of that used by fuller and leal to study laminar flows @xcite . for turbulence ,
the shear rate @xmath3 is a random variable .
the pcs method enables determination of the time - averaged shear rate @xmath0 , its standard deviation @xmath1 , and the gaussian transform of the probability density function ( pdf ) @xmath2 itself . because the method has not been used before , the values of the mean shear @xmath0 obtained by pcs are compared with those measured by laser doppler velocimetry ( ldv ) @xcite .
the pcs scheme has the advantage of improved signal - to - noise , short data - collection times , and also the compactness of the apparatus .
the pcs scheme can be used when the mean flow rate is absent or present .
hence it may be useful outside of the domain of turbulence studies . with the pcs scheme ,
a single beam illuminates a group of moving particles that scatter light into a photodetector at some scattering angle @xmath5 .
the inset of fig.[setup]a shows the incident and scattered laser beam of momentum @xmath6 and @xmath7 , respectively and the scattering vector @xmath8 at a point in the flowing soap film , an incident beam is focused to a bright spot of size @xmath4 . the intensity @xmath9 of the incident beam is taken to be gaussian form , @xmath10 figure [ setup]b , a side view of the setup ,
will be discussed below .
the velocity at any point @xmath11 can then be written as the velocity at the center of the spot @xmath11=0 plus a term proportional to the shear rate tensor @xmath12 , which is the quantity of interest .
the dominant component of @xmath13 near a wall in this experiment is @xmath14 , where @xmath15 is in the flow direction @xmath16 while @xmath17 is in the transverse direction in the film plane .
note that @xmath3 is a scalar quantity .
let @xmath18 be the velocity of an illuminated particle at a horizontal distance @xmath17 from the center of the incident beam ( @xmath19 ) .
then @xmath20 where the higher order terms have been neglected . within a multiplicative constant , the scattered electric field from @xmath21 particles within the incident beam at time @xmath22
is @xmath23 here @xmath24 is the incident gaussian field at the position of the @xmath25 particle . because the scattering from micron - size particles is almost perfectly elastic , @xmath26 where @xmath27 is the vacuum wavelength of the incident light beam ( 633 nm ) and @xmath28 is the refractive index of the soap film , which is 99 % water
. it will first be assumed that the flow is laminar , so that @xmath12 is time - independent , that is to say , the pdf of the shear tensor is a delta function centered at the mean value of @xmath12 . then the intensity correlation function which is simply related to the electric field autocorrelation function through the bloch - siegert theorem @xcite ( which is applicable to any gaussian pdf , including a delta function ) is @xmath29 where @xmath30 evaluating ( 3 ) , using ( 1)-(2 ) and averaging over @xmath22 gives a result previously obtained by fuller et al .
@xcite for laminar flow , as opposed to a turbulent one .
they evaluated @xmath31 rather than @xmath32 . in the experiments described below
, the turbulent soap film flows in the @xmath16 direction with mean velocity @xmath33 , where this average is over the width @xmath34 of the soap film . then @xmath35 where @xmath36 is an average over time .
use has been made here of the gaussian form of the incident beam . because @xmath3 is a random function for turbulent flows , an additional average over @xmath3
is needed , giving @xmath37 ( with @xmath2 having its maximum near @xmath0 ) ; @xmath31 is the gaussian transform of @xmath38 two important parameters , in addition to @xmath4 , are @xmath34 and the reynolds number , @xmath39 , where @xmath40 is the kinematic viscosity of the soap solution .
if the supporting walls that bound the film flow are not smooth , their roughness @xmath41 is another important control parameter . as in three dimensions one expects @xcite the dimensionless frictional drag @xmath42 to be independent of @xmath43 when it is sufficiently large .
it now depends on the ratio @xmath44 @xcite . at intermediate values of @xmath43 ,
experiment @xcite and theory @xcite support the result @xmath45 , where @xmath46 is just a number , and in 3d flows , @xmath47 if the shear tensor has more than one component @xcite @xmath48 with @xmath49 when the fluid is incompressible , as in this experiment @xcite .
the factors to the left of the integral take into account the extraneous effects of particle diffusion and transit time broadening ; they are discussed later in the text .
( a ) : setup for vertically flowing soap film .
the film flows down from reservoir rt through valve v between strips rw and sw , separated by width @xmath34 .
the weight w keeps the the nylon wires taut .
inset shows scattering diagram .
( b ) : side view of the setup , showing laser source , focusing lens and photodetector ] the experiments were performed on a soap film channel , shown in fig.[setup ] with @xmath34=2 cm . the flow is driven by gravity , but there is an appreciable opposing force from air friction . however , near the vertical plastic strips that support the film , the viscous force from the wires dominates @xcite .
these strips are glued to thin plastic wires 0.5 mm in diameter that join to form an inverted v at the top and at the bottom , as indicated in fig.[setup ] . at the apex
, a small tube connects the reservoir to a valve v that controls the flow rate .
the wires at the bottom connect and deliver the spent soap solution to reservoir rb , where it is pumped back to the top reservoir rt to keep the flow rate steady .
typical flow rates of the soap solution are @xmath50 0.2 ml / s . the soap solution is 1% dawn dishwashing detergent in water . it is loaded with neutrally buoyant polystyrene particles which scatter the incident beam from a 5 mw 633 nm he - ne laser into the photodetector , a perkin elmer spcm - aqr-12-fc .
the laser source is located behind the soap film while the photodetector is located in front of it as shown in fig.[setup]b .
the laser beam is focused onto the soap film with a lens of focal length 25 cm .
the photon stream is delivered to the photodetector through an optical fiber , where the fiber tip is located 7 cm from the illuminated spot on the film . in these experiments ,
@xmath4 is limited by the wavelength of visible light , focal length of the focusing lens and the diameter of the incident beam and has a value of @xmath4 = 100 @xmath51 m .
the scattering vector @xmath52 is in the vertical ( @xmath16 ) direction and the dominant component of the shear rate is @xmath53 the diameter @xmath54 of the seed particles ( 0.4 @xmath51 m ) is sufficiently small that their stokes number in the strongest turbulence is less than 0.1 @xcite
. hence the particle velocities are adequately close to that of the fluid .
the refractive index of the soap solution is roughly 1.3 .
typically , the scattering angle @xmath55 , @xmath56 m@xmath57 . using a seed - particle density of 1.5 gm / l yields an average photon counting rate of 10@xmath58 hz .
experiments were performed with a horizontally oriented comb penetrating the soap film at a point above the measuring point and with the comb absent . only with the comb present
is the turbulence reasonably developed and the energy spectrum is of scaling form , @xmath59 , with @xmath60 @xcite .
this is the @xmath61 , defined as the interval where vorticity of larger size fluctuations cascade to smaller scales . in two dimensions
there is also a cascade of energy fluctuations to larger scales , where @xmath62 = 5/3 , as in three dimensions .
however , it is not accessible for decaying turbulence , as in this experiment @xcite . by making the bounding walls rough , so that turbulence is constantly being generated there , the inverse energy cascade
can also be seen @xcite .
the teeth of the comb as well as their spacing is 2 mm . to further test the pcs technique ,
measurements are also made with the comb absent . in this case , there is no well - defined energy spectrum decaying as a power law .
nevertheless the flow is far from laminar , so that @xmath0 can be measured by both pcs and ldv .
to first order , @xmath63 and @xmath64 .
however , experimentally @xmath31 is found to be a non - gaussian function . if @xmath65 , @xmath66 which is clearly non - gaussian in @xmath67 .
both panels of fig .
[ corrwall ] show that while @xmath2 is close to gaussian form , the gaussian fit ( solid lines ) is not perfect .
these `` good '' fits to gaussian form were unexpected .
there are two other effects that can contribute to the decay of @xmath31 : thermal diffusion of the seed particles and transit time broadening , which can be dominant for large @xmath68 .
both of these contributions are small in these experiments but are easy to correct for @xcite . to take diffusion into account , one multiplies eq . ( 4 ) by the factor , @xmath69 where @xmath70 is the diffusion constant , which , for spherical particles of diameter @xmath54 is @xmath71 , where @xmath72 is boltzmann s constant and @xmath73 is viscosity . as for the transit time effect ,
particles passing through a beam of size @xmath4 produce a burst of light intensity that temporally modulates the scattered light .
the multiplicative correction factor here is @xmath74@xcite .
the decay times for both of these effects is long compared to the viscous decay time of interest so these multiplicative time factors can be dropped .
for example , with a spot size @xmath75 m and a typical mean velocity of @xmath33 = 1 m / sec , the transit time @xmath76 associated with the effect is or order @xmath77 0.1 ms .
this is fifty times longer than typically measured @xmath78 .
diffusion times are much longer than this and hence contribute insignificantly to the decay of @xmath31 .
fig.[corrwall ] shows @xmath31 for measurements made with the comb absent ( a ) and present ( b ) , respectively . here
@xmath79 2 m / s in both experiments , @xmath34=2 cm , and @xmath4 = 100 @xmath51 m .
the vertical axis is linear , but the horizontal axis is @xmath80 , so as to display several decades of lag time .
the insets to both figures show @xmath81 vs @xmath82 , so that a gaussian decay of @xmath31 appears as a straight line .
the straight lines in the lower insets indicate that @xmath31 is indeed of gaussian form for very small @xmath67 .
they are a best fit to the experimental curves and correspond @xmath0 = 1600 s@xmath57 and 1000 s@xmath57 for the experiments with and without the comb .
the solid lines in the upper insets to fig.[corrwall ] are best fits under the assumption of a gaussian @xmath38 a good fit clearly extends beyond the small-@xmath67 limit and enables the determination of the standard deviations @xmath1 of the mean shear as well as @xmath0 itself . the mean shear @xmath0 is calculated from the definition of variance , @xmath83 .
the results are @xmath0 = 950 hz @xmath1 = 300 hz with the comb absent and @xmath0 = 1620 hz @xmath1 = 500 hz with the comb present .
the ratio of @xmath1 to @xmath0 is near 20 % .
the shear measurement is done in the viscosity - dominated layer of width @xmath84 , where @xmath16 is the distance from the comb .
ideally the spot size @xmath4 should be much smaller than @xmath84 .
the function @xmath85 is proportional to @xmath17 within @xmath84 .
prior experiments have established that at @xmath16 = 20 cm below the comb , where the measurements were made , @xmath86 is roughly 200 @xmath51 m @xcite .
thus the beam size @xmath4 is small enough to correctly measure the viscous shear rate .
the single - point pcs measurements of @xmath0 are now compared with those of ldv , made in the traditional way ; the vertical velocity component @xmath15 is measured at two nearby horizontally - spaced points in the viscosity - dominated interval .
the ldv measurements were made 2.5 cm below the pcs beam spot , which is 80 cm below point p in fig.[setup ] and 20 cm below the comb .
the ldv measurement point is advanced in 50 @xmath51 m steps starting at @xmath17=0 .
the minimum useful value of @xmath17 is dictated by the necessity of avoiding strong light scattering from the supporting plastic strip with its edge at @xmath17=0 .
the ldv laser source is 514 nm line from a coherent argon - ion laser operating at a power of 500 mw , roughly one hundred times that used in the pcs device .
the data collection time for each measurement of @xmath85 is roughly 20 s. because the correlation time is of the order of microseconds , and the counting rate is of the order of mhz the function form of @xmath31 emerges after only a few seconds of data collection with the correlator .
the mean shear rate @xmath0 in the viscous region obtained by ldv and pcs agree to within one standard deviation , as seen in table [ tab : comb_shear_rate ] .
the uncertainties are deduced from seven measurements made at the indicated values of @xmath33 . from an individual run
, one can not extract @xmath1 from the ldv data , because noise fluctuations can change even the sign of @xmath87 .
the ldv and pcs measurements span the range 29000 @xmath8845000 and from 40000 @xmath8857000 , with and without comb respectively . with the comb in place , the taylor microscale reynolds number @xmath89 , where @xmath90 = 1 mm .
the errors from one run to another are _ not _ statistical in origin .
rather , the source is variations in the flow speed through the valve and the motion of the film plane caused by velocity fluctuations of the surrounding air which could be only partially suppressed by placing the entire apparatus in a tent .
.[tab : comb_shear_rate]mean shear rate @xmath0 as measured by ldv and pcs in a narrow range of mean flow speeds ( comb inserted ) . [
cols="^,^,^",options="header " , ] fig.[pcs_ldv_comp_15 ] shows measurements of @xmath0 as a function of @xmath17 in units of 50 @xmath51 m obtained using pcs ( circles ) and ldv ( triangles ) in the range out to @xmath17 = 1.50 mm with the comb present . here
@xmath33 = 2.16 m / s , @xmath34=2 cm and the kinematic viscosity of the soap solution is close to that of water ( @xmath40 = 0.01 @xmath91/s ) , @xmath43=45,000 .
the main messages conveyed by this graph are ( a ) the two schemes give roughly the same results for @xmath92 , ( b ) near one of the walls , the ldv measurements are noisier ( for reasons already discussed ) , and ( c ) @xmath0 decreases with increasing @xmath17 . even in the absence of air friction , this decrease is expected and is well - studied in 3d flows @xcite . in soap film
flows , air friction slows the flow far from the walls , making analysis of the data there difficult .
this experiment indicates it should be possible to measure @xmath0 near the wall and in the interior of 3d flows , though care must be taken to collect scattered photons from only a small volume in the fluid .
far from a bounding wall in 3d turbulence , the pcs method will suffer from the limitation that @xmath4 should be smaller than the smallest eddy size @xcite , defined as @xmath93 . even in these soap film experiments ,
@xmath73 is estimated to be comparable to or smaller than @xmath4 . yet , as fig.[pcs_ldv_comp_15 ] shows , the ldv measurements of @xmath0 agree with the pcs result up to @xmath17 = 1.5 mm from a wall , well outside the viscous region . as a function of distance from the wall ( in mm ) with a comb in place to strengthen the turbulence .
the mean flow speed @xmath33 = 2.16 m / s , @xmath43=45,000.,width=3 ]
though the photon correlation scheme has been used here to measure properties of the shear rate in a two - dimensional soap film , it can be used in three dimensional flows as well .
the pcs method has good signal - to - noise , is compact , and uses a laser in the mw range .
the method yields the variance of the shear rate as well as its mean value .
the correlation function itself is the gaussian transform of the probability density function , @xmath2 .
we wish to thank t. tran , t. adamo , w. troy , n. goldenfeld , n. guttenberg , and k. nguyen for their contributions to this work .
this work is supported by nsf grant no .
dmr 0604477 and nsf fellowship 60016281 to s. steers . | a photon correlation method is introduced for measuring components of the shear rate tensor in a turbulent soap film .
this new scheme , which is also applicable to three - dimensional flows , is shown to give the same results as laser doppler velocimetry , but with less statistical noise .
the technique yields the mean shear rate @xmath0 , its standard deviation @xmath1 , and a simple mathematical transform of the probability density function @xmath2 of the shear rate itself . 0.2 cm |
there has been considerable study of the bond percolation cluster considered as a random - resistor network , with each occupied bond having unit resistance and non - occupied bonds having infinite resistance @xcite . in two dimensions ,
the configuration studied is typically an @xmath3 lattice and the conductance is measured between two opposite sides which are assumed to have infinite conductance .
the backbone of the cluster is then defined as the set of bonds that are connected to the two sides having infinite conductance through paths that have no common bond . at the percolation threshold , the backbone mass scales as @xmath4 with @xmath5 @xcite and in this `` bus bar '' geometry
is strongly correlated with @xmath6 .
the total conductance of the backbone as a function of @xmath6 has been studied extensively and has been found to scale as @xmath7 with @xmath8 @xcite .
recently , the distribution of masses of backbones defined by two _ points _ ,
i.e. , backbones defined as the set of those bonds that are connected by paths having no common bonds to two points separated by distance @xmath1 within an @xmath3 lattice , has been studied @xcite .
one finds that when @xmath9 , there is a very broad distribution of backbone masses for a given @xmath1 .
figure [ fig1 ] illustrates some typical percolation clusters and their backbones defined in this configuration . because of the broad distribution of backbone masses we have the opportunity to study the conductance between these two points separated by a fixed distance , @xmath1 , as a function of the mass of the backbone defined by these points .
one might expect that , for fixed @xmath1 , the average conductance would increase with increasing backbone mass because there could be more paths through which current can flow .
in fact , we find that the average conductance decreases monotonically with increasing backbone size , in contrast with the behavior of homogeneous systems and non - random fractals in which conductance increases with increasing @xmath2 .
we explain our finding by noting that the conductance is strongly correlated with the shortest path between the two points , and then studying the distribution of shortest paths between the two points for a given @xmath2 .
this analysis extends recent studies of the distribution of shortest paths where no restriction on @xmath2 is placed @xcite .
our system is a two - dimensional square lattice of side @xmath10 with points @xmath11 and @xmath12 defined as @xmath13 , @xmath14 . for each realization of bond percolation on this lattice , if there is a path of connected bonds between @xmath11 and @xmath12 , we calculate ( i ) the length of the shortest path between @xmath11 and @xmath12 , ( ii ) the size of the backbone defined by @xmath11 and @xmath12 and ( iii ) the total conductance between @xmath11 and @xmath12 .
we obtain data from 100,000 realizations for each of 8 values of @xmath1 ( 1 , 2 , 4 , 8 , 16 , 32 , 64 , and 128 ) at the percolation threshold , @xmath15 . we bin these results based on the value of the backbone mass , @xmath2 by combining results for all realizations with @xmath16 and taking the center of each bin as the value of @xmath2 . in fig .
[ fig2](a ) , we plot the simulation results for the average conductance @xmath0 and find that the conductance , in fact , decreases with increasing @xmath2 .
the decrease is seen more clearly in fig .
[ fig2](b ) , in which we plot scaled values as discussed below .
in non - fractal systems , the conductance increases as the mass of the conductor increases .
we next consider the average conductance on the sierpinski gasket , a non - random fractal , the first three generations of which are illustrated in fig .
[ fig3](a)(c ) . because the sierpinksi gasket is not translationally invariant , the analogue of the average conductivity between two points in the percolation cluster is the conductivity averaged over all pairs of points separated by distance @xmath1 . at each successive generation , there are two types of pairs : ( i ) pairs which correspond to pairs in the previous generation ( e.g. , a and b ) and ( ii ) pairs which do not correspond to pairs in the previous generation ( e.g. , d and e ) .
it is obvious that as we move from one generation to the next , the conductance between pairs of type ( i ) increases because there are more paths between the points then in the previous generation . on the other hand ,
the conductance between the pairs of type ( ii ) are lower on average than between the pairs present in the previous generation because on average the shortest path between the two points is longer than between the pairs in the previous generation . however , for any given @xmath1 , the shortest path between any two points has a fixed upper bound independent of the generation . due to this bound on the shortest path ,
the net result is that the average conductivity increases with succeeding generations .
this is shown in fig .
[ fig3](d ) in which we plot the average conductivity calculated exactly for generations 1 to 6 for @xmath17 , 2 and 4 .
in order to understand why the average conductance of the percolation backbone decreases with increasing @xmath2 , we must ( i ) recognize that the conductance is strongly correlated with the shortest path between the two points and ( ii ) study @xmath18 , the distribution of shortest paths between the two points for a given backbone mass .
hence we next create the @xmath18 probability distribution , binning the data logarithmically by taking the average over samples centered at @xmath19 .
figure [ fig4](a ) shows the results of the simulations for @xmath18 for @xmath17 for various backbone masses .
the plots collapse , the only difference in the plots being the values of the upper cut - offs due to the finite size of the backbone .
figure [ fig1 ] illustrates how the size of the backbone constrains the possible values of the shortest path .
for all values of @xmath2 , a section of each plot in fig .
[ fig4](a ) exhibits power law behavior . in fig .
[ fig4](b ) , we show the distributions @xmath18 for different @xmath1 and a given @xmath2 . in fig .
[ fig4](c ) we see that when scaled with @xmath20 the plots collapse , so we can write @xmath18 in the scaling form @xmath21 an expression for @xmath22 can be found by recognizing that we can write the well - studied distribution @xmath23 , the probability that the shortest path between two points separated by euclidean distance @xmath1 is @xmath24 , independent of @xmath2 , as @xmath25 where ( i ) @xmath26 is the distribution of backbone masses given distance @xmath1 between the points which determine the backbone and ( ii ) @xmath27 is the lower cutoff on @xmath2 given @xmath24 .
@xmath26 has the form @xcite @xmath28\ ] ] where @xmath29 is the fractal dimension of the backbone and @xmath30 is the exponent for the blob size distribution @xcite . from ref .
@xcite @xmath31 where @xmath32 is the fractal dimension of the shortest path .
since @xmath33 and @xmath34 , implying @xmath35 , the lower cutoff @xmath27 in eq .
( [ f2 ] ) scales as @xmath36 as @xmath37 , the upper cutoff is @xmath38 because the maximum backbone mass is not constrained by the length of the shortest path . substituting eqs .
( [ f3 ] ) , ( [ f4 ] ) , and ( [ f4a ] ) into eq .
( [ f2 ] ) , and equating powers of @xmath1 ( or powers of @xmath24 ) of the left and right hand sides of the resulting equation , we find @xmath39 using @xmath30 and the values @xmath40 @xcite and @xmath41 @xcite , we find @xmath42 , in good agreement with our simulation result @xmath43
we can now calculate the average conductance .
since @xmath44 is strongly correlated with @xmath24 , and since @xmath44 scales with @xmath1 as @xmath45 and @xmath24 scales with @xmath1 as @xmath46 , we have @xmath47 then using the fact that @xmath48 we have @xmath49 where @xmath50 now @xmath18 is nonzero only for @xmath51 where @xmath52 and @xmath53 are constants . hence using @xmath54 , we find @xmath55 is nonzero for @xmath56 using these bounds to normalize the distribution , we find @xmath57 then @xmath58^{-\tilde\mu(z+2)}\over 1-\left[{(bm_b)^{1/d_b}\over ar}\right]^{-\tilde\mu(z+1)}}.\end{aligned}\ ] ] thus as @xmath2 goes to infinity , @xmath0 decreases asymptotically to a constant as @xmath59^{-\tilde\mu(z+1)}\right].\ ] ] by considering the asymptotic dependence of @xmath0 on @xmath2 , we can reasonably fit the simulation results by choosing the parameters @xmath52 and @xmath53 in eq .
( [ f9 ] ) to be 0.9 and 6 , respectively . using these values for @xmath52 and @xmath53 , in fig .
[ fig2](a ) we plot @xmath60 from eq .
( [ f9 ] ) for multiple values of @xmath1 and find that agreement with the simulation results improves with increasing @xmath1 . for large @xmath1 ,
the curves for the simulation results and the curves for the theoretical results are coincident at large @xmath2 .
the poor results for small @xmath1 are due to corrections - to - scaling not included in our derivation ( e.g. , for small @xmath1 , there are significant corrections - to - scaling for the relations @xmath61 and @xmath34 @xcite ) .
equation ( [ f9 ] ) can be re - cast in terms of the scaled variable @xmath62 as @xmath63 where @xmath64 in fig . [
fig2](b ) , we plot the average conductance scaled in accordance with eqs .
( [ f12 ] ) and ( [ f13 ] ) .
the expected collapse improves with increasing @xmath1 for the same reason as noted above . above the percolation threshold , for backbones of size larger than the correlation length ,
the strong correlation between the conductance and the shortest path breaks down and we expect the conductance to _ increase _ with the mass of the backbone , as is the case in non - random systems .
this is seen in fig .
[ fig2](c ) , where we plot conductance versus backbone mass for the bond occupation probabilities @xmath65 and @xmath66 , which are above the percolation threshold and , for comparison , conductance _ at _ the percolation threshold , @xmath67 @xcite .
figure [ fig2](c ) shows that for @xmath66 , all backbone masses sampled are of size greater than the correlation length and the conductance increases monotonically . for @xmath65 ,
the smaller backbone masses are of size less than the correlation length and fig .
[ fig2](c ) shows that the conductance initially decreases ; for larger backbone masses , however , the sizes of the backbones are greater than the correlation length and fig .
[ fig2](c ) shows that the conductance then increases .
the derivation of eq .
( [ f9 ] ) and its agreement with the results of our simulations confirm our understanding of why the average conductance decreases with increasing backbone mass : the smaller contributions to the average conductance from the longer minimal paths possible in the clusters with larger backbone size cause the average conductance to be smaller .
our derivation was not specific to two dimensions , and should also hold in higher dimensions .
we thank j. andrade for helpful discussions , and bp amoco for financial support . because , above the percolation threshold , the backbone mass is strongly correlated with the system size , we create backbones of a given mass by varying the system size .
the simulations of fig .
2(c ) for the plots above the percolation were obtained choosing @xmath68 , 10 , 20 , 40 , 80 , 160 , and 200 . | on two - dimensional percolation clusters at the percolation threshold , we study @xmath0 , the average conductance of the backbone , defined by two points separated by euclidean distance @xmath1 , of mass @xmath2 .
we find that with increasing @xmath2 and for fixed @xmath1 , @xmath0 asymptotically _ decreases _ to a constant , in contrast with the behavior of homogeneous sytems and non - random fractals ( such as the sierpinski gasket ) in which conductance increases with increasing @xmath2 .
we explain this behavior by studying the distribution of shortest paths between the two points on clusters with a given @xmath2 .
we also study the dependence of conductance on @xmath2 slightly above the percolation threshold . 2 |
accretion processes in astrophysical disks lead to enormous luminosity and huge changes in disk structure during astronomically short times .
efficient transport mechanisms are necessary to achieve such short time - scales .
anisotropic turbulence appears to be a major physical condition to provide astrophysical disks with strong transport .
as these disks generally exhibit increasing specific angular momentum towards larger radii and thus fulfill the stability criterion of rayleigh , they do not give rise to an instability leading to turbulence by themselves .
searches for instabilities in disks with rotation profiles similar to a keplerian one unveiled several ways to turbulence being more or less favorable with respect to their prerequisites for the disk configuration .
gravitational instability needs the disk to be either cool or massive .
nonlinear and nonaxisymmetric perturbations require a severe additional perturber near the disk ; conditions for instability in a purely hydrodynamical disk were derived by dubrulle ( 1993 ) .
hydrodynamic instability essentially comes down to violating the rayleigh criterion saying that a rotation profile with an increasing specific angular momentum with radius is hydrodynamically stable , that is for @xmath5 . since the typical length of perturbations caused be external forces
is supposed to be very large , the required amplitude of the perturbations has to be considerable , too , in order to violate the rayleigh criterion locally .
if the length scale of the perturbation is comparable to the radius , a strong alteration of the keplerian velocity of several per cent is needed .
convection was shown to deliver either negligible transport ( stone & balbus 1996 ) or inward angular momentum transport ( kley et al . 1993 ) .
the requirements for the magnetic shear - flow instability ( balbus & hawley 1991 ) do match astrophysical conditions in accretion disks in many configurations . all it needs is a radially decreasing angular velocity and a weak magnetic field threading the rotating object . it can even be shown that the temperature range applicable to the magnetic shear - flow concept is very broad ; even very small ionization fractions are sufficient to magnetize a disk in many cases ( balbus & hawley 1998 ) .
first numerical approaches to the magnetic shear - flow instability tackled the local problem ; the linear analysis as described by balbus & hawley ( 1991 ) were immediately followed by 2d simulations ( hawley & balbus 1991 ) of a small box - shaped domain which was cut out of the disk .
these computations confirmed the relation between magnetic - field strength and wavenumber derived from the linear analysis earlier , they showed that the maximum growth rate of any wavenumber is independent from the field strength and that the system is capable of transporting angular momentum . because of being restricted to axisymmetric problems
, they could not provide self - sustained turbulence which needs dynamo - action , and the slow decay of the turbulence is an indirect effect of the cowling theorem . improved computations dealt with a three - dimensional even though local configuration , and particular care was taken for the radial boundary conditions which are not simply periodic , but account for the shear due to differential rotation ( hawley et al .
these comprehensive computations indeed resulted in magnetically sustained turbulence whose anisotropy causes efficient outward angular momentum transport .
this work was followed up by stratified models ( stone et al .
1996 ) covering more than 50 orbital periods of the box cut - out . as the computational domain covered 2 density scale heights
, this work was a first step towards global simulations , followed up by similar approaches such as ziegler & rdiger ( 2000a , b ) .
linear studies of global configurations of disks threaded by magnetic fields in various directions were carried out .
curry & pudritz ( 1995 ) investigated the stability for vertical and azimuthal fields threading the disk .
they found in detail that the actual initial field geometry does not strongly depend on field topology as was suggested by the numerical simulations of hawley & balbus ( 1991 ) .
rdiger et al .
( 1999 ) particularly addressed the angular momentum transport in their linear study .
these investigations are now followed by nonlinear simulations of a compressible fluid with density stratification in global computational domains .
the most recent global approach was presented by hawley ( 2000 ) following the evolution of a thick torus under the influence of an external magnetic field threading parts of the computational domain .
the magnetic shear - flow instability was found to set in quickly causing enough turbulent viscosity to soon form a keplerian velocity profile .
all the above mentioned numerical studies integrate the ideal mhd equations omitting magnetic diffusivity .
local simulations including diffusivity were performed by fleming et al .
( 2000 ) which proved the onset of instability even for low conductivity .
the full understanding of accretion disks implies self - excited dynamo action as well as angular momentum transport .
can positive angular momentum transport be coupled with a suitable kinetic helicity providing the expected dynamo action according to the @xmath2-effect principle ? as the kinetic helicity in stratified , rotating disks is expected to be negative , a wrong sign ( positive ) would follow for the dynamo-@xmath2-effect according to the conventional @xmath2-theories .
we present global three - dimensional diffusive simulations and study the angular momentum transport as well as dynamo action and the sign of @xmath6 as a consequence of the correlation with the flow .
the present computations are not meant as particular simulations of a star formation process .
they focus on the applicability of the magnetic shear - flow instability for an astronomically fast process like the formation of a star .
such global simulations are still a challenge for modern computers and fast algorithms .
the computations presented here make use of the zeus-3d code developed for astrophysical problems of magnetohydrodynamics ( see the key papers stone & norman 1992a , b ; stone et al .
1992 for numerical , clarke et al .
1994 for technical details ) .
we use cylindrical coordinates and an extensive computational domain covering radii @xmath7 to 6 and @xmath8 to 6 , a vertical extension of @xmath9 to + 1 , and the full azimuthal range of @xmath10 to @xmath11 . see table [ tab1 ] for a list of models presented in this paper . in this approach , we assume an isothermal disk to save computation time on the energy equation . the remaining system for integration is @xmath12 where @xmath13 , @xmath14 , and @xmath15 are the density , velocity , and magnetic field resp . ; @xmath16 is the pressure ( @xmath17 in our model ) , @xmath18 is the gravitational potential ( solely from a central mass @xmath19 ) , @xmath20 is the current density , and @xmath21 is the magnetic diffusivity which is not an original ingredient to the zeus code .
it is constant in time and space .
the computations of the electromotive force in the routines emfs , mocemfs , and hsmoc of zeus-3d are extended with the appropriate @xmath22 components .
we chose @xmath23 to 0.01 .
the additional time - step criterion resulting from the diffusivity is roughly 0.1 for the finest grid used here .
it is therefore irrelevant for the determination of the time - step which is typically @xmath24 or , in the case of strong fluctuations of velocity and magnetic field , one or two orders of magnitude lower . in physical units , this diffusivity is large and in fact accounting for a subgrid turbulent diffusivity which can not be resolved . [ cols="^,<,<,^,^,^,<",options="header " , ] the distribution of @xmath1 versus time and the @xmath25-direction is shown in figure [ alptime_bgauss ] for models v. outward angular momentum transport is marked with white areas , inward transport by dark areas .
strong transport is observed at large distances from the equator shortly after the onset of the instability .
regions of outward transport migrate towards the equator .
this is particularly marked in the model - v plot for which we found a development into a strong - transport regime ( cf .
figure [ plotalpha_bgauss ] ) .
positive angular momentum transport is not restricted to high altitudes anymore .
power spectra resulting from fourier transforms of the magnetic field of model v over the the azimuthal and radial directions are shown in figures [ fourcolor_bgauss35 ] and [ fourcolor_bgauss35r ] , respectively .
the azimuthal decomposition of the velocity field is shown in fig .
[ fourcolor_bgauss35v ] .
the individual azimuthal spectra for all points in the ( @xmath26)-plane were averaged , as were the points of the ( @xmath27)-plane for the radial spectra . in the azimuthal spectra
, we observe slight maxima in power at @xmath28 and @xmath29 , followed by a strong , roughly power - law decline between @xmath30 and @xmath31 .
the magnetic spectrum is steeper than the kinetic one .
the underlying power - laws would be @xmath32 to @xmath33 which is significantly steeper than a kolmogorov spectrum , and is similar to what armitage ( 1998 ) found for a global simulation in the far-@xmath34 range .
the kolmogorov spectrum of isotropic turbulence would exhibit a wavenumber exponent of @xmath35 ; here we face mhd turbulence which implies an effect of alfvn waves impeding the transfer of energy towards smaller scales . the energy spectrum of isotropic mhd turbulence is @xmath36 , whence more shallow than the kolmogorov spectrum . contrasting with the azimuthal decomposition , the radial spectrum matches the theoretical mhd spectrum over the entire range of modes as can be seen in fig .
[ fourcolor_bgauss35r ] . choosing the rough dimensions of an inner protostellar system of 100 au radius for scaling the mass flux through our models , we obtain an accretion rate in physical units as shown in fig
. [ plotoutflux_cs10bh ] .
widely accepted values for the accretion rates of protostellar disks around solar - type stars are @xmath37 to @xmath38/yr .
the compilation of results for those stars by strom ( 1994 ) rather favors @xmath38/yr . the accretion rate increases with stella mass and may already exceed @xmath39/yr for a 3-solar - mass star .
our results such as given in fig .
[ plotoutflux_cs10bh ] are thus not too far from the accretion rates deduced from the observations . among stratified local simulations ,
we find a number of very similar results as regards the magnitude of @xmath1 , such as in stone et al .
( 1996 ) who obtained @xmath40 , hawley et al .
( 1996 ) where an average of 0.016 was found , with typical values of 0.01 .
the largest box used delivered 0.02 , the smallest , 0.003 . upon studying the dependence of the turbulence on the rotation law @xmath41 , hawley et al .
( 1999 ) found @xmath42 for @xmath43 .
a most recent local - box simulation by ziegler & rdiger ( 2000b ) covering 288 orbital periods resulted in a space - time average of 0.015 , where the reynolds part was @xmath44 and the maxwell part was @xmath45 .
the stress was normalized with the averaged equatorial - plane pressure and thus provides a lower limit for @xmath1 .
the early stratified - box simulations by brandenburg et al .
( 1995 ) provided @xmath1 an order of magnitude lower between 0.001 and 0.005 .
all the models agree upon the dominance of maxwell stresses over reynolds stresses .
it is most interesting to note that other global simulation such as performed by armitage ( 1998 ) or hawley ( 2000 ) result in an order of magnitude higher values of 0.2 to 0.3 .
contrasting with this , the global spherical , but unstratified simulation by drecker ( 2000 ) gives again an @xmath1 which is an order of magnitude lower than shown here , i.e. @xmath46 .
those simulations alone included a physical viscosity .
the simulations of hawley ( 2000 ) and with somewhat different objectives , those of machida et al .
( 2000 ) had open boundaries in radial and vertical directions .
this fact limits the run - time of the models which are prone to running out of matter .
the long - term existence of non - zero toroidal fields in each hemisphere as shown for example in figure [ bphi_bgauss ] for model v indicates the possible action of a dynamo in the disk . the local - box simulations by hawley et al .
( 1996 ) impose severe constraints on the growth of an average magnetic field because of the periodicity of the boundary conditions .
a non - zero term of the average magnetic - field growth is due to the shear and requires the presence of an average radial component , in the local nomenclature @xmath47 .
we follow their derivation of the averages by replacing the volume curl integral by a vector - produc surface integral of ( 3 ) .
the average @xmath48 is equivalent to @xmath49 and delivers average components as @xmath50 these considerations show that slight initial fluctuations in @xmath51 are able to create non - zero average magnetic fields at any time , even if their @xmath52 vanishes , as a consequence of shear and radial accretion through @xmath53 . in particular , the non - constant vertical profile of the orbital velocity produces a @xmath54 from the initial perturbation @xmath55 .
fluctuations in the accretion flow generate @xmath56 and @xmath55 .
the actual radial differential rotation then quickly produces @xmath54 from even very small @xmath56 , and toroidal fields are prone to dominate the disk fields .
kinematic dynamos have been extensively examined in numerous previous publications concerning various types of geometries . a widely used mean - field approach
to circumnavigate the difficulties with resolving the small scales in a model applies the likely amplification of a magnetic field through a fluctuative electromotive force parallel to the actual large - scale magnetic field .
this effect is generally expressed by the term ` @xmath2-effect ' and can be thoroughly studied in e.g. krause & rdler ( 1981 ) . in a model
splitting large and small scales , the induction equation extends beyond the large - scale electromotive force such as @xmath57 the @xmath2-effect is part of the development of the small - scale electromotive force as @xmath58 where @xmath59 and @xmath60 are tensors in general .
if written in components , the induction equation shows that the essential component of the @xmath6 tensor is the @xmath61 component which converts azimuthal fields into radial and vertical fields , whereas the differential rotation converts radial and vertical fields back into the toroidal component by the large - scale part of @xmath62 .
this approximation is usually referred to as an @xmath63-dynamo .
second - order correlation approximation leads to a relation of @xmath6 with the helicity of the flow , @xmath64 which is , when approximated by a typical time - scale @xmath65 , @xmath66 the same principle is applicable to the current helicity ( keinigs 1983 ) giving @xmath67 among other issues , the evolution of these quantities for the helicity will be subject of the following sections .
special attention is payed to the symmetry of the magnetic fields of the solutions . in general
, the preference of a certain symmetry is a tool to check the consistency with mean - field dynamo models .
additionally , dipolar ( antisymmetric ) fields are supposed to support the launch of winds forming jets from the disk better than quadrupolar ( symmetric ) fields , with the latter rather providing closed field lines within the disk .
the mean - field disk dynamo of rekowski et al .
( 2000 ) produced dipolar magnetic fields for negative @xmath6 in the northern hemisphere .
a number of mean - field dynamo models in one to three dimensions delivered varying symmetry of the solutions with a major influence of the boundary conditions vacuum versus perfectly conducting .
three - dimensional investigations of stability by meinel et al .
( 1990 ) and elstner et al .
( 1992 ) found quadrupolar solutions for the condition of low - conductivity surroundings of the disk , whereas perfectly conducting boundaries delivered dipolar solutions . a local model with periodic boundaries
will thus not be able to answer the question of the final symmetry of the solutions .
an ideal global model would comprise the entire investigated object and is less depending on the surroundings ( which will be vacuum in very good approximation ) .
the parity of a the toroidal magnetic field is measured by @xmath68 where @xmath69 and @xmath70are the symmetric and antisymmetric energy parts resp . , and
@xmath71 is the total energy of in toroidal field component .
figures [ parityphd_cs10bh ] and [ parityphd_bgauss ] show the temporal development of parity of the toroidal magnetic fields in the models ii and v along with a repeated plot of @xmath1 .
we should emphasize that the initial perturbation of the magnetic field had dipolar symmetry .
since the total flux through the vertical surfaces , however , vanishes , field configurations with either symmetry may arise from the evolution of the model .
the dipolar initial perturbation may not have been fully re - organized as the diffusive time - scale @xmath72 is 1600 rotation periods for the low - diffusivity model ii and 160 revolutions for the high - diffusivity model v. the latter covered @xmath73 .
an additional experiment used a completely mixed initial magnetic - field perturbation according to a distribution which has zero parity in @xmath55 , and the toroidal component immediately emerging from the slight vertical has zero parity , too .
again , the magnetic flux through the vertical surfaces is zero .
such initial conditions have been applied to models ii and v. the runs lasted roughly 10 revolution resulting in parity variations with a slight tendency towards quadrupolar ( symmetric ) solution ( @xmath74 from + 0.1 to + 0.2 ) .
average kinetic helicities were computed based on a layer slightly below one density scale - height in both hemispheres . the average is taken in @xmath75- and @xmath76-direction .
there is a considerable scatter in the time series of the helicity , but temporal averages are all negative on the northern side for all models .
the only exception is the northern part of model ii with a near - zero value .
negative sign also holds for the two test models with mixed - parity initial perturbation .
average current helicities are also negative on the northern side ( and positive on the southern one ) throughout almost all models
. the only exceptions are the mixed - parity models which show a current helicity with opposite sign compared with the kinetic helicity in both hemispheres .
since the angular momentum transport is dominated by magnetic stresses , we also ask about the _ energy _ in the flow fluctuations and the magnetic field .
the temporal evolutions of these two quantities in figures [ helres_paper_cs10bh ] and [ helres_paper_bgauss ] show a clear difference between the models ii and v. kinetic - energy dominance holds for model ii , whereas magnetic dominance is found for model v. in order to evaluate the applicability of the @xmath6 approach , we compute the average toroidal field @xmath77 and compare it with the actual average electromotive force @xmath78 derived directly from the simulated vector fields .
a correlation between the two quantities may justify the @xmath2-effect for dynamo generation of magnetic fields .
as a meaningful @xmath6 must change its sign at the equator , we plot the correlation for both hemispheres of the disk separately ; the result is given in figure [ dyn_temp2 ] .
as the temporal fluctuations are strong , we plot the time - averages of @xmath54 and @xmath79 in figure [ emf_bgauss ] only with their resulting sign .
the quantities are also averaged in azimuthal direction and plotted in the @xmath80-plane .
white areas denote a negative sign ; the dominance of negative averaged electromotive forces indicates a positive @xmath6 in the northern hemisphere in accordance with figure [ dyn_temp2 ] .
the classical understanding of the coriolis force giving preference to left - handed helicity on the northern hemisphere ( whence positive @xmath6 ) appears not applicable for the correlation plots of models ii and iii .
the fluctuations were strong , and no meaningful sign of @xmath6 can be derived .
however , models ia , v , and vi show a mostly negative averaged emf for both hemispheres , while @xmath77 changes its sign .
an indication for positive @xmath6 is thus found from these simulations .
in fact , models v and vi are those which show saturated kinetic energies .
the other significant difference to the indefinite models ii and iii is the 10times higher diffusivity , @xmath81 .
notwithstanding , an indication for a negative @xmath6 was already found in simulations of a local box cut out of the disk by brandenburg et al . ( 1995 ) .
the same sign is found by ziegler & rdiger ( 2000b ) from long runs of local simulations , although a clear influence of resistivity on the evolution was found there .
lowest magnetic reynolds numbers are @xmath82 for model v in this paper , calculated with the velocity difference between inner and outer radial boundary , just as ziegler & rdiger did for the local box .
these considerations are somewhat limited , since a straight - forward regression line in fig .
[ dyn_temp2 ] will have a significant offset from the origin of the graph ; a vanishing @xmath77 does apparently not coincide with a vanishing @xmath83 .
the effect of the other field components is considered small though ; while the average hemispherical @xmath77 are of the order of 100 , the @xmath84 is of order 10 and @xmath85 of order unity .
a considerable contribution is suspected from the current density through the @xmath60-term in ( [ emf ] ) , and is appears to be quite natural that an offset in the correlation graphs in figure [ dyn_temp2 ] is found . for this reason
, we did not plot such regression lines in the figure .
additionally , the sole consideration of the induction equation in the `` classical '' dynamo theory leads to the omission of the lorentz force , which is not only essential for the onset of the magneto - rotational instability but for the maintenance of the turbulence and the transport of angular momentum as well .
it is thus obvious that the neglect of the back - reaction of magnetic fields on the flow need not lead to representative models of keplerian disks .
three - dimensional global simulations have shown , that the magnetic shear - flow instability is a fast mechanism to generate a turbulent flow in a keplerian disk .
we presented models with a computational domain covering the full azimuthal range , accounting for accretion of matter .
the shakura - sunyaev parameter @xmath1 increases rapidly during the first revolutions after switching on the magnetic field and reaches @xmath42@xmath86 at this stage of the computations , though it appears not yet fully saturated in several of the models .
maxwell stresses exceed reynolds stresses almost entirely .
the latter undergo strong fluctuations and are partly negative .
indications for a dynamo action in the disk are found , the corresponding dynamo-@xmath6 tending to be negative north of the equatorial plane and positive south of the equatorial plane .
the generated magnetic fields may maintain the turbulence even if the external field ceases .
the strong excitation of low - order azimuthal modes in the magnetic field is another promising fact for dynamo action with respect to the cowling theorem . for various reasons , a representative example for an accretion disk dynamo
is provided by model v ( and nearly as well by the similar model vi with higher inflow velocity ) .
the time series is sufficiently long covering more than 18 revolution periods .
the angular momentum transport reaches a saturated level of high efficiency during the last four orbital periods .
the magnetic diffusivity is roughly two orders of magnitude higher than the numerical diffusivity .
the striking dominance of one sign of the averaged emf in figure [ emf_bgauss ] indicates a physically evolved , relevant state of the system .
the results connected with the outcomes of mean - field dynamo theory include include the following facts : ( i ) a negative average toroidal magnetic field is found for the northern hemisphere , a positive on the southern one .
the averaged emf is negative in both hemisphere indicating a _ positive _ @xmath6-effect on the northern side ( negative on the southern side ) .
( ii ) the correlation plot of average toroidal field and emf gives the same picture .
( iii ) _ negative _ kinetic and magnetic helicities on the northern hemisphere ( positive on the southern one ) are also consistent with a positive @xmath6 .
( iv ) the _ dipolar structure _ of the solution contrasts with the results from mean - field dynamo models of disks which yield quadrupolar solutions for positive @xmath6 . we have to add though that the disks of our paper are _ not thin_. mean - field simulations by covas et al .
( 1999 ) show a transition from quadrupolar to dipolar fields when the opening angle of the disk is enlarged , rather representing a torus than a disk .
the extension of the run - times of such simulations exceeding the order of diffusion times promises further results about dynamo action in accretion disks .
the connection with mean - field concepts are most interesting as well as the implications for jet - launch models . the invaluable help by d. elstner , potsdam , in adapting the zeus-3d code for our purposes is greatly acknowledged .
we are grateful to the john v. neumann - institut for computing at the forschungszentrum jlich , germany , for the opportunity to use the cray t90 computer .
thanks for the kind support by the deutsche forschungsgemeinschaft .
armitage , p.j .
, 1998 , apj 501 , l189 balbus , s.a . , hawley , j.f . , 1991 , apj 376 , 214 balbus , s.a . , hawley , j.f . , 1998 , rev
70 , 1 balbus , s.a . ,
hawley , j.f . ,
stone , j.m .
, 1996 , apj 467 , 76 brandenburg , a. , nordlund , . ,
stein , r.f .
, torkelsson , u. , 1995 , apj 446 , 741 clarke d.a .
, norman m.l . , fiedler r.a . , 1994 , zeus-3d user manual , univ . of illinois covas , e. , tavakol r. , tworkowski , a. , brandenburg , a. , brooke , j. , moss , d. , 1999 , a&a 345 , 669 curry , c. , pudritz , r.e . , 1995 ,
apj 453 , 697 drecker , a. , rdiger , g. , hollerbach , r. , 2000 , mnras 317 , 45 dubrulle , b. , 1993 , icarus 106 , 59 elstner , d. , meinel , r. , beck , r. , 1992 , a&a suppl .
94 , 587 evans , c. , hawley , j.f . , 1988 ,
apj 33 , 659 fleming , t.p . ,
stone , j.m . ,
hawley , j.f . , 2000 ,
apj 530 , 464 hawley , j.f . , 2000 ,
apj 528 , 462 hawley , j.f . , balbus , s.a . ,
1991 , apj 376 , 223 hawley , j.f . , balbus , s.a . ,
winters , w.f . , 1999 , apj 518 , 394 hawley , j.f . ,
gammie , c.f .
, balbus , s.a . , 1995 ,
apj 440 , 742 hawley , j.f . , gammie , c.f .
, balbus , s.a .
, 1996 , apj 464 , 690 keinigs , r.k . , 1983 ,
fluids 26 , 2558 kley , w. , papaloizou , j.c.b . ,
lin , d.n.c .
, 1993 , apj 416 , 679 krause f. , rdler k .- h .
mean - field magneto - hydrodynamics and dynamo theory .
machida , m. , hayashi , m.r . ,
matsumoto , r. , 2000 , apj 532 , l67 meinel , r. , elstner , d. , rdiger , g. , 1990 , a&a 236 , l33 rekowski , m.v . ,
rdiger , g. , elstner , d. , 2000 , a&a 353 , 813 rdiger , g. , primavera , l. , arlt , r. , elstner , d. 1999 , mnras 308 , 887 shakura , n.i . , sunyaev , r.a . , 1973 ,
a&a 24 , 337 stone j.m .
, balbus s.a . , 1996 , apj 464 , 346 stone j.m . ,
norman m.l . , 1992a , apj suppl . 80 , 753 stone j.m . , norman m.l .
, 1992b , apj suppl .
80 , 791 stone j.m . , mihalas d. , norman m.l . , 1992 , apj suppl .
80 , 819 stone j.m . ,
hawley j.f . , gammie , c.f .
, balbus , s.a . , 1996 ,
apj 463 , 656 strom s.e , 1994 , rev .
astrofs . 29 , 23 ziegler , u. , rdiger , g. , ( 2000a ) , a&a 356 , 1141 ziegler , u. , rdiger , g. , ( 2000b ) , a&a , in press | we perform global three - dimensional simulations of accretion disks integrating the compressible , non - viscous , but diffusive mhd equations .
the disk is supposed to be isothermal .
we make use of the zeus-3d code integrating the mhd equations and added magnetic diffusivity .
we measure the efficiency of the angular - momentum transport . various model simulations delivered transport parameters of @xmath0 to 0.05 which are consistent with several local numerical investigations .
two of the models reach a highly turbulent state at which @xmath1 is of the order of 0.1 . after a certain stage of saturating of the turbulence
, reynolds stress is found to be negative ( inward transport ) in many of the models , whereas maxwell stresses dominate and deliver a positive ( outward ) total transport .
several of the models yield strongly fluctuating reynolds stresses , while maxwell stresses are smooth and always transport outwards .
dynamo action is found in the accretion disk simulations .
a positive dynamo-@xmath2 is indicated in the northern hemisphere of the most prominent run , coming along with negative kinetic and current helicities ( all having the opposite sign on the southern side ) .
the dipolar structure of the magnetic field is maintained throughout the simulations , although indication for a decay of antisymmetry is found .
the simulations covered relatively thick disks , and results of thin - disk dynamo models showing quadrupolar fields may not be compatible with the results presented here .
@xmath3 @xmath4 a&a |
despite recent observational progress , the 40 years old gamma - ray bursts ( grbs ) mystery is far from being completely solved ( see e.g. @xcite for a review ) .
there is general consensus on the cosmological nature of these transient sources of gamma - ray radiation , and , at least for the long burst category , the association with the explosion of massive stars ( > 30 m@xmath1 ) is a scenario that reproduces most observed features , although the number of grbs spectroscopically associated with supernovae is still limited .
whatever the exact progenitors of long grbs are , they have been detected up to redshift 6.7 @xcite , making them powerful tools to investigate the early universe ( star formation history , re - ionization , etc . ) , and to possibly derive cosmological parameters . for short bursts , on the other hand ,
the situation is less clear , mainly because of the lack of a statistically compelling number of good quality observations of their afterglows .
models involving the coalescence of two compact objects ( black holes , neutron stars or white dwarfs ) are however commonly invoked to explain these events , as their average distance is smaller with respect to long bursts .
many questions concerning grbs are still open , such as the physical processes at work during the prompt phase , in terms of particle acceleration and radiation processes , the grbs classification , a better characterization of grb host galaxies and progenitors , as well as some fundamental physics issues like lorentz invariance , the origin of cosmic rays and gravitational waves . in order to contribute to address the above questions ,
the french space agency ( cnes ) , the chinese academy of sciences ( cas ) and the chinese space agency ( cnsa ) are developing the _ svom _ mission ( space - based multi - band astronomical variable object monitor ) .
_ svom _ has successfully reached the end of its phase a design study , and is planned to be launched in 2013 in a circular orbit with an inclination of @xmath030@xmath2 and altitude of @xmath0600 km .
it will carry four instruments : eclairs , a coded mask wide field telescope that will provide real time localizations of grb to arcminute level , two grms units , non - imaging gamma - ray spectrometers , and two narrow - field instruments , xiao and vt , for arcsecond localizations , and for the study of the early afterglow phases in the x - ray and optical bands . indeed ,
once a grb is detected within field of view of eclairs , the satellite will autonomously perform a slew towards the grb direction , in order to allow the observations of the afterglow by the xiao and vt telescopes . a possible realization of the _ svom _ satellite is sketched in fig .
[ fig : svom ] .
the _ svom _
pointing strategy derives from a combination of two main constraints : the avoidance of bright x - ray galactic sources and an anti - solar pointing , to have the grbs always detected on the night side of the earth .
even if the latter choice induces some dead time at mission level , due to the earth passages occulting eclairs field of view once per orbit , it will enhance the possibility of successful follow - up with large ground based facilities , with a goal of 75% of _ svom _ grbs easily observable during their early afterglow phase .
besides the space flown instruments , the _ svom _ mission includes a set of ground based instruments , in order to broaden the wavelength coverage of the prompt and of the afterglow phase : gwacs are a set of wide field optical cameras that cover a large fraction of eclairs field of view .
they will be based in china and will follow eclairs pointings , in order to catch the prompt optical emission associated with grbs . two robotic telescopes ( gfts ) ,
one based in china , and one provided by cnes , complete the ground based instrumentation .
their goal is to measure the photometric properties of the early afterglow from the near infra - red to the optical band , and to refine the afterglow position provided by the on - board instruments .
a summary of the multi - wavelength capabilities of svom is reported in fig .[fig : svom ] . in the following sections the instruments and the mission
are described in some detail .
eclairs is made of a coded mask telescope working in the 4250 kev energy range ( cxg ) , and a real - time data - processing electronic system ( uts , see below ) .
the cxg has a wide field of view ( @xmath0 2 sr ) , and a fair localization accuracy ( @xmath010 arcmin error radius ( 90% c.l . ) for the faintest sources , down to a couple of arcmin for the brightest ones ) .
its detector plane is made of 80@xmath380 cdte pixels yielding a geometrical area of 1024 @xmath4 .
the telescope is passively shielded , and a new generation electronics developed at cea saclay together with the careful detector selection , and the optimized hybridization done at cesr toulose allow to lower the detection threshold with respect to former cdte detectors by about 10 kev , reaching @xmath04 kev .
the cxg , in spite of its rather small geometrical surface , is thus more sensitive to grbs with soft spectra , potentially the most distant ones , than currently flying telescopes .
a scheme of the cxg telescope is shown in left panel of fig .
[ fig : eclairs ] , while the right panel of the same figure shows how the cxg predicted sensitivity compares to current and past missions as a function of the grb peak energy .
detection assuming a band @xcite spectrum with the other spectral as parameters reported in the plot.,title="fig : " ] detection assuming a band @xcite spectrum with the other spectral as parameters reported in the plot.,title="fig : " ] the eclairs / cxg telescope is expected to localize about 70 grbs per year .
this estimate takes into account the dead time induced by the passages over the southern atlantic anomaly , that increase significantly the instrumental particle induced background , and by the passage of the earth in the cxg field of view .
the uts ( unit de traitement scientifique ) is in charge of analyzing eclairs data stream in real - time and of detecting and localizing the grbs occurring within its field of view .
it will implement two separate triggering algorithms , one based on the detection of excesses in the detectors count rate , and using imaging as trigger confirmation and localization , and a second one based on imaging , that is better suited for long , slowly rising grbs .
once a grb is detected , its position will be sent to the ground via a vhf emitter antenna . at the same time
the grb position will be transmitted to the platform that may autonomously slew to the grb direction in three to five minutes in order to bring it in the field of view of the narrow field instruments , xiao and vt .
the uts will also collect relevant information in the cxg and grm data stream , such as light curves and sub - images , and send them to the earth via the vhf channel , to promptly provide additional information on the trigger quality and on the grb spectral and temporal characteristics . for more details on eclairs , see @xcite and @xcite .
the gamma - ray monitor ( grm ) on board _ svom _ is composed of two identical units each made of a phoswich ( nai / csi ) detector of 280 @xmath4 , read by a photomultiplier . in front of each detector
there is a collimator in order to reduce the background and to match the cxg and grm fields of view . a schematic view of one grm unit and the combined cxg / grm sensitivity are shown in fig . [ fig : grm ] .
the grm does not have imaging capabilities , however as can be seen from fig .
[ fig : grm ] , the grm extends the spectral coverage of the _ svom _ satellite to the mev range .
this is an important point , since the current detectors like bat on board _ swift _ @xcite or ibis / isgri on board _ integral _ @xcite have a comparable or better localization accuracy with respect to the cxg , but they lack a broad band coverage , hampering a correct spectral characterization of the prompt high - energy emission of grbs , which is a key input of any sensible modeling of grbs radiative processes .
the recent successful launch of _ fermi _ , and the availability of the gbm detectors ( non - imaging spectrometers with a 2@xmath5 field of view ) used in synergy with bat and isgri will partially fill this need before the launch of _ svom _ , but the different orbits , pointing constrains , and sensitivities of the three instruments imply a low rate of simultaneous detections . on the other hand , for every _ svom _ grb a good localization and good spectral information will be available at the same time .
detection . a band spectrum with @xmath6=1 , @xmath7=2.5 , @xmath8=100 kev , and f@xmath9=1 photon @xmath10 s@xmath11
is overplotted for comparison.,title="fig : " ] detection .
a band spectrum with @xmath6=1 , @xmath7=2.5 , @xmath8=100 kev , and f@xmath9=1 photon @xmath10 s@xmath11 is overplotted for comparison.,title="fig : " ]
the x - ray imager for afterglow observations ( xiao ) will be provided by an italian consortium , lead by the inaf - iasf institute in milan .
it is a focusing x - ray telescope , based on the grazing incidence ( wolter-1 ) technique .
it has a short focal length of @xmath00.8 m , and a field of view ( 25 arcmin diameter ) adequate to cover the whole error region provided by the cxg telescope , so that after the satellite slew the grb position should always be inside the xiao field of view .
xiao has an effective area of about 120 @xmath4 and the mirrors are coupled to a very compact , low noise , fast read out ccd camera , sensitive in the 0.52 kev energy range .
the sensitivity of the xiao telescope is reported in fig .
[ fig : xiao ] , and simulations based on a sample of light curves collected with xrt on board _ swift _ show that virtually all grb x - ray afterglows are detectable by xiao during the first hours .
this means in practice that each grb , for which a satellite slew is performed ( not all grb can be pointed due to different constraints at platform level ) , can be localized with a @xmath0 5 arcsec accuracy , see fig .
[ fig : xiao ] . indeed
the source localization accuracy , is linked to the number of detected photons as @xmath12 where k is a constant depending on the instrument point spread function , and n is the number of detected photons , and the early afterglows will provide enough photons to reach the degree of positional accuracy mentioned above . for more details on the xiao instrument ,
see @xcite detection in 10 ks ) , computed assuming a 30@xmath13 half energy diameter , a source extraction circle of 15@xmath13 radius , and an energy bin @xmath14e = e/2 .
right : xiao location accuracy as a function of the number of detected photons . for each afterglow
at least 100 counts are expected to be collected by xiao.,title="fig : " ] detection in 10 ks ) , computed assuming a 30@xmath13 half energy diameter , a source extraction circle of 15@xmath13 radius , and an energy bin @xmath14e = e/2 .
right : xiao location accuracy as a function of the number of detected photons . for each afterglow
at least 100 counts are expected to be collected by xiao.,title="fig : " ]
the space - borne visible telescope will be able to improve the grb localizations obtained by the cxg and xiao to sub - arcsecond precision through the observation of the optical afterglow .
in addition it will provide a deep and uniform light - curve sample of the detected optical afterglows , and allow to do primary selection of optically dark grbs and high - redshift grb candidates ( z>4 ) .
the field of view of the telescope will be 21@xmath321 arcmin , sufficient to cover the error box of the cxg .
the detecting area of the ccd has 2048 x 2048 pixels to ensure the sub - arcsecond localization of detected sources .
the aperture of the telescope should guarantee a limiting magnitude of @xmath15 = 23 ( 5@xmath16 ) for a 300 s exposure time .
such a sensitivity is a significant improvement over the uvot on board the _ swift _ satellite ( see fig .
[ fig : vt ] ) and over existing ground - based robotic grb follow - up telescopes .
the vt is expected to detect nearly 70% of svom grbs for which a slew is performed .
the telescope will have at least two bands in order to select high - redshift grb candidates .
they are separated at 650 nm , which corresponds to a redshift of z@xmath04 - 4.5 using ly@xmath6 absorption as the redshift indicator .
the ground segment of the mission will be composed of x- and s - band antennas ( for data and housekeeping telemetry download ) , a mission operation center , based in china , two science centers ( based in china ( csc ) and france ( fsc ) and in charge of operations and monitoring of the scientific payload ) , and a vhf alert network.the latter will be composed of a series of receivers distributed over the globe in order to guarantee continuous coverage for the alerts dispatched by the platform .
the alerts will contain the information about the grb positions , that will be sent to the ground as soon as more accurate information is derived on board , followed by complementary quality indicators ( light curves , images , etc . )
produced on board .
the vhf network is directly connected to the fsc , which is in charge of formatting and dispatching the alerts to the scientific community through the internet ( gcns , vo events , svom web page , etc . ) .
the first alerts corresponding to the initial localization by the cxg are expected to reach the recipients one minute after the position has been derived on board .
then the following alerts , containing the x - ray position of the afterglow , and eventually the optical position computed from vt data , will reach the community within 10 minutes from the first notice . in case of a refined ( sub - arcsec ) position
is available from the prompt data analysis of the gfts or gwacs ( see below ) , this information will immediately reach the the fsc , and dispatched to the scientific community through the channels mentioned above . for more details on the alert distribution strategy ,
see @xcite .
in addition the fsc will be on charge of publishing the cxg pointing direction in order to facilitate ground based robotic telescopes to quickly react , minimizing the slew time .
the ground - based wide - angle camera array is designed to observe the visible emission of more than 20% of _ svom _ grbs from 5 minutes before to 15 minutes after the grb onset .
the array is expected to have an assembled field of view of about 8000 deg@xmath17 and a 5@xmath16 limiting magnitude @xmath15 = 15 for a 15 s exposure time for full moon nights . to comply with both the science requirements and technical feasibility , each camera unit will have an aperture size of 15 cm , a 2048 x 2048 ccd and a field of view of 60 deg@xmath17 . in total about 128 camera units
are required to cover the 8000 deg@xmath17 .
two svom robotic telescopes will automatically position their 20 - 30 arcmin field - of - view to the position of grb alerts and , in case of a detection , they will determine the position of the source with 0.5 arcsec accuracy .
both telescopes will be provided with multi - band optical cameras and the french gft will also have a near infra - red ccd . since one telescope can only observe those candidates occurring above the horizon and during night at the telescope site , these sites will be located in tropical zones and at longitudes separated by 120@xmath2 at least , in order to fulfill the requirement of a 40% efficiency .
these telescopes could also be used to follow alerts which are not considered to be reliable enough to be distributed to the whole community .
this procedure allows increasing the chance of detecting low s / n events , while not wasting the observing time of instruments which are outside the _ svom _ collaboration .
the scientific objectives of the gfts include the quick identification and characterization of interesting grbs ( e.g. highly redshifted grbs , whose visible emission is absorbed by the lyman alpha cutoff and the lyman alpha forest , dark bursts , nearby grbs ) , and multi - wavelength follow - up of 40% of _ svom _ grbs ( at optical and x - ray wavelengths ) from 30 to 10@xmath18 seconds after the trigger .
this will be done with the gft and the vt at optical wavelengths and cxg and the xiao at x - ray wavelengths .
this will allow measuring the spectral energy distribution of the burst during the critical transition between the grb and the afterglow . d.g . acknowledges the french space agency ( cnes ) for financial support .
j.d . & y.q .
are supported by nsfc ( no . 10673014 ) .
s.m . acknowledges the support of the italian space agency through contract i/022/08/0 . 9 piran , t. 2005 , reviews of modern physics , 76 , 1143 fynbo , j. p. u. , greiner , j. , kruehler , t. , et al .
2008 , grb coordinates network , 8225 , 1 band , d. , matteson , j. , ford , l. , et al .
1993 , apj , 413 , 281 s. schanne , 2008 , proceedings of the santa fe conference .
aip conference proceedings , volume 1000 , 58 remou , n. , barret , d. , godet , o. , et al .
2009 , these proceedings gehrels , n. , chincarini , g. , giommi , p. , et al .
2004 , apj , 611 , 1005 lebrun , f. , leray , j. p. , lavocat , p. , et al .
2003 , a&a , 411 , l141 mereghetti , s. , de luca , a. , fiorini , m. , et al . 2008 , spie meeting , arxiv:0807.0893v1 akerlof , c.v . , & swan , h.f .
2007 , apj , 671 , 1868 claret , a. , 2008 , proceedings of the santa fe conference .
aip conference proceedings , volume 1000 , 569 | we present the _ svom _
( space - based multi - band astronomical variable object monitor ) mission , that is being developed in cooperation between the chinese national space agency ( cnsa ) , the chinese academy of science ( cas ) and the french space agency ( cnes ) , and is expected to be launched in 2013 .
its scientific objectives include the study of the grb phenomenon ( diversity and unity ) , grb physics ( particle acceleration , radiation mechanisms ) , grb progenitors , cosmology ( host galaxies , intervening medium , star formation history , re - ionization , cosmological parameters ) , and fundamental physics ( origin of cosmic rays , lorentz invariance , gravitational waves sources ) .
_ svom _ is designed to detect all known types of gamma - ray bursts ( grbs ) , to provide fast and reliable grb positions , to measure the broadband spectral characteristics and temporal properties of the grb prompt emission .
this will be obtained in first place thanks to a set of four space flown instruments . a wide field ( @xmath02 sr ) coded mask telescope ( eclairs ) , operating in the 4250 kev energy range ,
will provide the triggers and localizations , while a gamma - ray non - imaging spectrometer ( grm ) , sensitive in the 50 kev-5 mev domain , will extend the prompt emission energy coverage .
after a satellite slew , in order to place the grb direction within field of view of the two narrow field instruments - a soft x - ray ( xiao ) , and a visible telescope ( vt ) - the grb position will be refined and the study of the early phases of the grb afterglow will be possible .
a set of three ground based dedicated instruments , two robotic telescopes ( gfts ) and a wide angle optical monitor ( gwac ) , will complement the space borne instruments .
thanks to the low energy trigger threshold ( @xmath04 kev ) of the eclairs , _ svom _ is ideally suited for the detection of soft , hence potentially most distant , grbs .
its observing strategy is optimized to facilitate follow - up observations from the largest ground based facilities .
address = cea saclay - dsm / irfu / service dastrophysique - orme des merisiers , f-91191 , gif - sur - yvette , france address = cea saclay - dsm / irfu / service dastrophysique - orme des merisiers , f-91191 , gif - sur - yvette , france , altaddress = astroparticules et cosmologie ( apc ) - 10 , rue alice domon et lonie duquet , f-75205 , paris cedex 13 , france address = laboratoire dastrophysique de marseille / cnrs / universit de provence , technopole de letoile , 38 rue frederic joliot - curie , 13388 marseille cedex 13 , france address = national astronomical observatories , chinese academy of sciences , 20a datun road , chaoyang district , beijing 100012 , china address = department of physics and center for astrophysics , tsinghua university , beijing 100084 , china address = laboratoire dastrophysique de toulouse - tarbes , universit de toulouse , cnrs , 14 av .
edouard belin , 31400 toulouse , france address = centre detude spatiale des rayonnements , universit de toulouse , centre nationale de la recherche scientifique , 9 avenue du colonel roche , 31028 toulouse cedex 04 , france address = cea saclay - dsm / irfu / service dastrophysique - orme des merisiers , f-91191 , gif - sur - yvette , france address = cea saclay - dsm / irfu / service dastrophysique - orme des merisiers , f-91191 , gif - sur - yvette , france address = national astronomical observatories , chinese academy of sciences , 20a datun road , chaoyang district , beijing 100012 , china address = xian institute of optics and precision mechanics , chinese academy of sciences , 17 xinxi road , xian , china address = national astronomical observatories , chinese academy of sciences , 20a datun road , chaoyang district , beijing 100012 , china address = national astronomical observatories , chinese academy of sciences , 20a datun road , chaoyang district , beijing 100012 , china address = centre detude spatiale des rayonnements , universit de toulouse , centre nationale de la recherche scientifique , 9 avenue du colonel roche , 31028 toulouse cedex 04 , france address = inaf istituto di astrofisica spaziale e fisica cosmica ( iasf ) , via bassini 23 , i-20133 , milano , italy address = national astronomical observatories , chinese academy of sciences , 20a datun road , chaoyang district , beijing 100012 , china address = key laboratory for particle astrophysics , institute of high energy physics , chinese academy of sciences , p.o.box 918 , 100049 , china |
in recent years , several new excited charmed baryon states have been observed by the babar , belle and cleo collaborations , such as the @xmath9 , @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 @xcite , and re - vivified the interest in the charmed baryon spectrum . on the other hand ,
the qcd sum rules is a powerful theoretical tool in studying the ground state heavy baryon states @xcite . in the qcd sum rules ,
the operator product expansion is used to expand the time - ordered currents into a series of quark and gluon condensates which parameterize the long distance properties of the qcd vacuum .
based on the quark - hadron duality , we can obtain copious information about the hadronic parameters at the phenomenological side @xcite . there have been several works on the masses of the heavy baryon states with the full qcd sum rules and the qcd sum rules in the heavy quark effective theory @xcite . in refs.@xcite
, we study the @xmath17 heavy baryon states @xmath18 , @xmath19 and @xmath20 and @xmath21 heavy baryon states @xmath22 , @xmath23 and @xmath24 with the full qcd sum rules , and observe that the pole residues of the @xmath21 heavy baryon states from the sum rules with different tensor structures are consistent with each other , while the pole residues of the @xmath17 heavy baryon states from the sum rules with different tensor structures differ from each other greatly .
those pole residues are important parameters in studying the radiative decays @xmath25 , @xmath26 and @xmath27 @xcite , we should refine those parameters to improve the predictive ability . in ref.@xcite ,
jido et al introduce a novel approach based on the qcd sum rules to separate the contributions of the negative - parity light flavor baryon states from the positive - parity light flavor baryon states , as the interpolating currents may have non - vanishing couplings to both the negative- and positive - parity baryon states @xcite .
before the work of jido et al , bagan et al take the infinite mass limit for the heavy quarks to separate the contributions of the positive and negative parity heavy baryon states unambiguously @xcite . in ref.@xcite
, we follow ref.@xcite and re - study the masses and pole residues of the @xmath17 flavor sextet heavy baryon states @xmath18 , @xmath19 and @xmath20 by subtracting the contributions of the negative parity heavy baryon states with the full qcd sum rules . in this article
, we use the same approach to study the @xmath28 flavor antitriplet heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 .
the article is arranged as follows : we derive the qcd sum rules for the masses and the pole residues of the heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 in sect.2 ; in sect.3 , we present the numerical results and discussions ; and sect.4 is reserved for our conclusions .
the @xmath31 flavor antitriplet heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 can be interpolated by the following currents @xmath32 and @xmath33 respectively , @xmath34 where the @xmath35 represents the heavy quarks @xmath36 and @xmath37 , the @xmath38 , @xmath39 and @xmath40 are color indexes , and the @xmath41 is the charge conjunction matrix .
the corresponding negative - parity heavy baryon states can be interpolated by the currents @xmath42 because multiplying @xmath43 to the @xmath44 changes the parity of the @xmath44 @xcite , where the @xmath44 denotes the currents @xmath32 and @xmath33 .
the correlation functions are defined by @xmath45 and can be decomposed as @xmath46 due to lorentz covariance .
the currents @xmath47 couple to both the positive- and negative - parity baryon states @xcite , i.e. @xmath48 , where the @xmath49 denote the negative parity baryon states .
we insert a complete set of intermediate baryon states with the same quantum numbers as the current operators @xmath50 and @xmath51 into the correlation functions @xmath52 to obtain the hadronic representation @xcite . after isolating the pole terms of the lowest states , we obtain the following result @xcite : @xmath53 where the @xmath54 are the masses of the lowest states with parity @xmath55 respectively , and the @xmath56 are the corresponding pole residues ( or couplings ) . if we take @xmath57 , then @xmath58 where @xmath59 \ , , \nonumber \\
b(p_{0 } ) & = & \pm{1 \over 2 } \left [ \lambda_+^{2 } \delta(p_{0 } - m_+ ) - \lambda_-^{2 } \delta(p_{0 } - m_{-})\right ] \ , , \end{aligned}\ ] ] the contribution @xmath60 ( @xmath61 ) contains contributions from the positive parity ( negative parity ) states only for the @xmath62 , while the contribution @xmath63 ( @xmath61 ) contains contributions from the negative parity ( positive parity ) states only for the @xmath64 .
we calculate the light quark parts of the correlation functions @xmath52 in the coordinate space and use the momentum space expression for the heavy quark propagators , then resort to the fourier integral to transform the light quark parts into the momentum space in @xmath65 dimensions , take @xmath57 , and use the dispersion relation to obtain the spectral densities @xmath66 and @xmath67 ( which correspond to the tensor structures @xmath68 and @xmath69 respectively ) at the level of quark - gluon degrees of freedom , finally we introduce the weight functions @xmath70 $ ] , @xmath71 $ ] , and obtain the following sum rules , @xmath72&=&\int_{\delta}^{\sqrt{s_0}}dp_0 \left[\rho^a(p_0 ) + \rho^b(p_0)\right]\exp\left[-\frac{p_0 ^ 2}{t^2}\right ] \ , , \end{aligned}\ ] ] @xmath73&=&\int_{\delta}^{\sqrt{s_0}}dp_0 \left[\rho^a(p_0 ) + \rho^b(p_0)\right]p_0 ^ 2\exp\left[-\frac{p_0 ^ 2}{t^2}\right ] \ , , \end{aligned}\ ] ] where @xmath74}{16\pi^2}\int_{t_i}^1 dt t\nonumber\\ & & + \frac{p_0}{128\pi^2}\langle \frac{\alpha_sgg}{\pi}\rangle \int_{t_i}^1 dt t-\frac{m_q^2}{768\pi^2}\langle \frac{\alpha_sgg}{\pi}\rangle \int_{t_i}^1 dt \frac{(1-t)^2}{t^2 } \delta ( p_0-\widetilde{m}_q)\nonumber \\ & & + \frac{m_s\left[3\langle\bar{q}g_s\sigma gq\rangle-\langle\bar{s}g_s\sigma gs\rangle \right]}{192\pi^2}\delta ( p_0-m_q ) + \frac { \langle\bar{q}q\rangle\langle\bar{s}s\rangle}{12}\delta(p_0-m_q ) \
, , \end{aligned}\ ] ] @xmath75}{16\pi^2}\int_{t_i}^1 dt \nonumber\\ & & + \frac{m_q}{128\pi^2}\langle \frac{\alpha_sgg}{\pi}\rangle \int_{t_i}^1 dt + \frac{m_q}{192\pi^2}\langle \frac{\alpha_sgg}{\pi}\rangle \int_{t_i}^1 dt \frac{(1-t)^3}{t^2}\nonumber \\ & & -\frac{m_q}{768\pi^2}\langle \frac{\alpha_sgg}{\pi}\rangle \int_{t_i}^1 dt \frac{(1-t)^2}{t}\widetilde{m}_q \delta(p_0-\widetilde{m}_q)\nonumber \\ & & + \frac{m_s\left[3\langle\bar{q}g_s\sigma gq\rangle-\langle\bar{s}g_s\sigma gs\rangle \right]}{192\pi^2}\delta ( p_0-m_q ) + \frac { \langle\bar{q}q\rangle\langle\bar{s}s\rangle}{12}\delta(p_0-m_q ) \ , , \end{aligned}\ ] ] @xmath76 @xmath77 where @xmath78 , @xmath79 , the @xmath80 are the threshold parameters , @xmath81 is the borel parameter , @xmath82 and @xmath83 in the channels @xmath30 and @xmath29 respectively .
the input parameters are taken to be the standard values @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 @xcite , @xmath89 @xcite , @xmath90 , @xmath91 and @xmath92 @xcite at the energy scale @xmath93 . the value of the gluon condensate @xmath94 has been updated from time to time , and changes greatly @xcite . at the present case ,
the gluon condensate makes tiny contribution , the updated value @xmath95 @xcite and the standard value @xmath89 @xcite lead to a difference less than @xmath96 for the masses . in the conventional qcd sum rules
@xcite , there are two criteria ( pole dominance and convergence of the operator product expansion ) for choosing the borel parameter @xmath81 and threshold parameter @xmath80 .
we impose the two criteria on the heavy baryon states to choose the borel parameter @xmath81 and threshold parameter @xmath80 , the values are shown in table 1 . from table 1 , we can see that the contribution from the perturbative term is dominant , the operator product expansion is convergent certainly . in this article , we take the contribution from the pole term is larger than @xmath97 ( @xmath98 ) for the positive ( negative ) parity baryon states , the uncertainty of the threshold parameter is @xmath99 , and the borel window is @xmath100 . in calculation , we neglect the contributions from the perturbative corrections . those perturbative corrections can be taken into account in the leading logarithmic approximations through anomalous dimension factors .
after the borel transform , the effects of those corrections are to multiply each term on the operator product expansion side by the factor , @xmath101^{2\gamma_{j}-\gamma_{\mathcal { o}_n } } $ ] , where the @xmath102 is the anomalous dimension of the interpolating current @xmath103 and the @xmath104 is the anomalous dimension of the local operator @xmath105 .
we carry out the operator product expansion at a special energy scale @xmath106 , and set the factor @xmath107^{2\gamma_{j}-\gamma_{\mathcal { o}_n}}\approx1 $ ] , such an approximation maybe result in some scale dependence and weaken the prediction ability . in this article
, we study the flavor antitriplet @xmath108 heavy baryon states systemically , the predictions are still robust as we take the analogous criteria in those sum rules . taking into account all uncertainties of the relevant parameters ,
we obtain the values of the masses and pole residues of the flavor antitriplet @xmath108 heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 , which are shown in figs.1 - 2 and table 2 . in this article
, we calculate the uncertainties @xmath109 with the formula @xmath110 where the @xmath111 denote the hadron mass @xmath112 and the pole residue @xmath113 , the @xmath114 denote the input qcd parameters @xmath115 , @xmath116 , @xmath117 , @xmath118 , @xmath119 , and the threshold parameter @xmath80 and borel parameter @xmath81 . as the partial derivatives @xmath120 are difficult to carry out analytically , we take the approximation @xmath121 ^ 2 $ ] in the numerical calculations . .
the borel parameters @xmath81 and threshold parameters @xmath80 for the heavy baryon states , the `` pole '' stands for the contribution from the pole term , and the `` perturbative '' stands for the contribution from the perturbative term in the operator product expansion . [ cols="^,^,^,^,^,^",options="header " , ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . , title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively . ,
title="fig:",width=188 ] of the heavy baryon states , the @xmath122 , @xmath123 , @xmath41 and @xmath65 correspond to the @xmath17 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively , while the @xmath128 , @xmath129 , @xmath130 and @xmath131 correspond to the @xmath132 heavy baryon channels @xmath124 , @xmath125 , @xmath126 and @xmath127 respectively .
, title="fig:",width=188 ] in ref.@xcite , bagan et al study the masses and couplings ( pole residues ) of the heavy baryon states @xmath133 , @xmath134 , @xmath126 and @xmath124 with the full qcd sum rules , and observe that the @xmath135 and @xmath136 have degenerate masses within the uncertainties , while the @xmath137 is lighter than the @xmath136 . in refs.@xcite ,
zhang et al perform a systematic study of the masses of charmed and bottom baryon states with the full qcd sum rules , where the vacuum condensates up to dimension six are taken into account ; while in ref.@xcite , duraes et al study only the masses of the heavy baryon states @xmath125 , @xmath127 and @xmath138 . in ref.@xcite ,
liu et al perform a systematic study of the masses of the bottom baryon states up to @xmath139 in the heavy quark effective field theory using the qcd sum rules . in the heavy quark limit , @xmath140 .
we can calculate the bound energy @xmath141 and the @xmath139 corrections of the lowest heavy baryon states @xmath29 with the qcd sum rules , compare them with the experimental data @xmath142 , and determine the heavy quark masses @xmath83 , which always suffer from large uncertainties , then use the @xmath83 as basic input parameters to calculate other heavy baryon masses , one can consult refs.@xcite for example . in table 3
, we also present the predictions from the full qcd sum rules and the qcd sum rules in the heavy quark effective theory ( where systematic studies are preformed ) @xcite . from tables 2 - 3
, we can see that the present and other theoretical predictions are all in good agreement with the experimental data for the positive parity baryon states .
the negative parity baryon states @xmath126 and @xmath127 are not observed yet , we make reasonable predictions for their masses to confront with experimental data in the future at the lhcb @xcite . the fractions @xmath143\exp\left[-\frac{p_0 ^ 2}{t^2}\right ] } { \int_{\delta}^{\sqrt{s_0}}dp_0\left[\rho^a(p_0 ) + \rho^b(p_0)\right]\exp\left[-\frac{p_0 ^ 2}{t^2}\right]}\end{aligned}\ ] ] are less than @xmath144 and @xmath145 in the positive ( negative ) parity charmed baryon and bottom baryon channels , respectively .
so the contaminations from the negative ( or positive ) parity baryon states are very small . in refs.@xcite
, we study the @xmath17 flavor sextet heavy baryon states @xmath18 , @xmath19 and @xmath20 with the full qcd sum rules , and observe that the pole residues from the sum rules with different tensor structures ( @xmath146 and @xmath69 ) differ from each other greatly . in ref.@xcite , we re - study the masses and pole residues of the @xmath17 flavor sextet heavy baryon states @xmath18 , @xmath19 and @xmath20 by subtracting the contributions of the negative parity heavy baryon states , and find that the predictions for the masses and pole residues are improved considerably . in the present case , we can choose the tensor structures @xmath146 or @xmath69 or @xmath147 freely to study the masses and pole residues .
the pole residues of the @xmath148 and @xmath149 heavy baryon sextets @xmath150 and @xmath151 have been calculated in our previous works @xcite .
once reasonable values of the pole residues @xmath7 and @xmath8 of the @xmath148 heavy baryon antitriplet @xmath152 are obtained , we can take them as basic input parameters and study the strong decays @xmath153 , @xmath154 and @xmath155 and the radiative decays @xmath156 and @xmath157 in a systematic ways with the light - cone qcd sum rules or the qcd sum rules in external field , and confront the predictions with the experimental data in the future at the besiii , @xmath158 and lhcb @xcite .
the strong decays @xmath159 and @xmath160 saturate approximately the widths of the @xmath161 and @xmath162 respectively , while the strong decays @xmath163 are seen @xcite . from our previous works @xcite , we can see that the corresponding strong decays @xmath164 , @xmath165 and @xmath166 are kinematically allowed ; although the bottom baryon states @xmath167 have not been observed experimentally yet . in refs.@xcite ,
we perform systematic studies for the radiative decays @xmath168 with the light - cone qcd sum rules as the strong decays @xmath169 are forbidden due to the unavailable phase space , while the radiative channels are not phase space suppressed and become relevant ; although the electromagnetic strength is weaker than that of the strong interaction .
the radiative decays @xmath170 and @xmath171 are important processes in testing the heavy quark symmetry and the chiral symmetry , for example , the @xmath172 and @xmath173 are governed by the radiative decays .
in this article , we study the @xmath28 flavor antitriplet heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 by subtracting the contributions from the corresponding @xmath174 heavy baryon states with the qcd sum rules , obtain the masses which are in good agreement with the experimental data and make reasonable predictions for the unobserved @xmath6 bottom baryon states . in calculation , we observe that the contaminations from the negative ( or positive ) parity baryon states are very small , one can choose the tensor structures @xmath146 or @xmath69 or @xmath147 freely to study the masses and pole residues .
once reasonable values of the pole residues @xmath7 and @xmath8 are obtained , we can take them as basic input parameters and study the strong decays and radiative decays in a systematic ways with the light - cone qcd sum rules or the qcd sum rules in external field , and confront the predictions with the experimental data in the future at the besiii , @xmath158 and lhcb .
this work is supported by national natural science foundation , grant number 10775051 , and program for new century excellent talents in university , grant number ncet-07 - 0282 , and the fundamental research funds for the central universities . | in this article , we study the masses and pole residues of the @xmath0 flavor antitriplet heavy baryon states ( @xmath1 , @xmath2 and ( @xmath3 , @xmath4 by subtracting the contributions from the corresponding @xmath5 heavy baryon states with the qcd sum rules , and observe the masses are in good agreement with the experimental data and make reasonable predictions for the unobserved @xmath6 bottom baryon states .
once reasonable values of the pole residues @xmath7 and @xmath8 are obtained , we can take them as basic parameters to study the revelent hadronic processes with the qcd sum rules .
+ zhi - gang wang + department of physics , north china electric power university , baoding 071003 , p. r. china pacs number : 14.20.lq , 14.20.mr key words : heavy baryon states , qcd sum rules |
it has long been a problem that the @xmath3 @xmath4 amplitude calculated in quenched lattice qcd is about a factor two too large compared to the experimental value@xcite . in this article
we report a new study of this problem , incorporating various theoretical and technical advances in recent years for analysis .
in particular we discuss in detail how one - loop corrections of chiral perturbation theory ( chpt ) recently calculated@xcite affect physical predictions for the decay amplitude from lattice qcd simulations .
our simulation is carried out in quenched lattice qcd employing the standard plaquette action for gluons at @xmath1 and the wilson action for quarks .
two lattice sizes , @xmath5 and @xmath6 , are employed .
we take up , down and strange quarks to be degenerate , and make measurements at four values of the common hopping parameter , @xmath7 , @xmath8 , @xmath9 and @xmath10 , which correspond to @xmath11 , @xmath12 , @xmath13 and @xmath14 .
the 4-quark operator most relevant for the @xmath3 @xmath2 decay is @xmath15 / 2 $ ] .
we extract the decay amplitude from the 4-point correlation function defined by @xmath16 where @xmath17 are wall sources for @xmath18 , @xmath19 and @xmath20 . in our calculations on a lattice of a temporal size @xmath21 , the walls are placed at the time slices @xmath22 , @xmath23 and @xmath24 .
the mesons are all created at rest , and the 4-quark operator @xmath25 is projected to zero spatial momentum .
we can use two types of factors to remove the normalization factors in @xmath26 .
if we define @xmath27 we find @xmath28 @xmath29 @xmath30 and @xmath31 @xmath32 for @xmath33 , where @xmath34 is the mass shift of the 2-pion state due to finite lattice size effects @xcite . in fig .
[ rwfrpf ] we plot @xmath35 and @xmath36 at @xmath37 as a function of time @xmath38 of the weak operator .
we clearly observe a non - vanishing slope for @xmath39 , while @xmath40 exhibits a plateau as expected .
the decay amplitude can be obtained by fitting @xmath39 to a single exponential or @xmath40 to a constant .
we find the results to be mutually consistent within the statistical error .
= 7.0 cm = 7.0 cm the one - loop renormalization factor relating the operator @xmath25 on the lattice and that in the continuum was obtained in refs .
we set @xmath41 or @xmath42 as the matching point and employ tadpole - improved perturbation theory with the mean - field improved @xmath43 coupling constant@xcite .
quark fields are normalized with the klm factor @xcite , @xmath44 .
in earlier calculations tree - level chiral perturbation theory ( chpt ) formula was used to obtain the physical amplitude @xmath45 from the lattice counterpart @xmath46 .
the formula takes the form @xmath47 where @xmath48 and @xmath49 are physical masses , and @xmath50 is the degenerate mass of @xmath51 and @xmath52 mesons on the lattice . clearly the physical amplitude should be independent of the lattice mass @xmath50 if the matching formula is exact . in fig .
[ comparison ] we compare results for the physical decay amplitude from a previous work@xcite carried out at @xmath53 on a @xmath54 spatial lattice with those of our simulation at @xmath1 .
our values obtained with the conventional quark normalization @xmath55 ( crosses and open squares ) are consistent with their results , but are larger than the experimental result roughly by a factor two .
let us also note with our results ( circles and squares ) that ( i ) the klm normalization and tadpole improvement of the operator have a significant effect on the amplitude , ( ii ) there is a clear dependence of the amplitude on the lattice meson mass @xmath50 , and ( iii ) a significant finite - size effect is observed between the two lattice sizes .
these features show that the tree - level chpt is inadequate to extract the physical amplitude from lattice results . in fig .
[ frho ] we show how predictions for the physical amplitude change if we apply the one - loop formula of chpt @xcite to our results . here
@xmath56 and @xmath57 are the cutoffs of chpt for quenched and full theory . in converting to physical values
we ignore effects of @xmath58 contact terms of the chpt lagrangian since their values are not well known .
an interesting point is that a size dependence seen with the tree - level analysis in fig .
[ comparison ] is absent .
another important point is that the magnitude of the amplitude decreases by @xmath59% over the range of meson mass covered in our simulation .
the amplitude depends significantly on the quenched lattice cutoff @xmath56 , in particular toward larger values of @xmath50 .
= 7.0 cm our results in fig .
[ frho ] show that a sizable meson mass dependence still remains .
this may be attributed to @xmath58 contact terms which were not taken into account .
if we assume that the contact terms of the full theory are as small as suggested by the phenomenological analysis @xcite , we can estimate the physical amplitude by a linear extrapolation of @xmath60 to the chiral limit . in table
[ finalfit ] we list results for several choices of the cutoff and the operator matching point @xmath61 . as is expected from fig . [ frho ] , variation of results on the choice of the cutoff parameters is small , being of the order of 10% .
allowing for this uncertainty , the values in table [ finalfit ] are consistent with the experimental value of @xmath62 .
the present study has shown that the quenched result for the @xmath0 decay amplitude agrees with experiment much better than previously thought , especially if the chiral limit is taken for the lattice meson mass . away from the chiral limit ,
the amplitude depends significantly on the cutoff scales of chpt , however .
this problem is largely ascribed to a mismatch of chiral logarithms between the quenched and full theories@xcite . in order to obtain @xmath0 decay amplitude free from this uncertainty ,
we need simulations in full qcd .
this work is supported by the supercomputer project ( no . 97 - 15 ) of high energy accelerator research organization ( kek ) , and also in part by the grants - in - aid of the ministry of education ( nos . 08640349 , 08640350 , 08640404 , 09246206 , 09304029 , 09740226 ) . c. bernerd and a. soni , nucl .
phys . * b*(proc . suppl.)*17 * ( 1990 ) 495 .
golterman and k. c. leung , hep - lat/9702015 .
m. lscher , comm .
* 105 * ( 1986 ) 153 .
g. martinelli , phys .
141b * ( 1984 ) 395 . c. bernard , t. draper , and a. soni , phys
* d36 * ( 1987 ) 3224 .
lepage and p.b .
mackenzie , phys .
* d48 * ( 1993 ) 2250 .
lepage , nucl . phys .
* b26 * ( proc.suppl . ) ( 1992 ) 45 ; p. mackenzie , nucl . phys . * b34 * ( proc.suppl . ) ( 1994 ) 35 ; a. kronfeld , nucl . phys .
* b34 * ( proc.suppl . ) ( 1994 ) 415 .
j. bijnens , h. sonoda and m. b. wise , phys .
* 53 * ( 1984 ) 2367 ; j. kambor , j. missimer and d. wyler , phys . lett . *
b261 * ( 1991 ) 496 . | a new study is reported of a lattice qcd calculation of the @xmath0 decay amplitude with the wilson quark action in the quenched approximation at @xmath1 .
the amplitude is extracted from the @xmath2 green function , and a conversion to the continuum value is made employing a recent one - loop calculation of chiral perturbation theory .
the result is consistent with the experimental value if extrapolated to the chiral limit . |
the notion of thin position for a closed orientable 3-manifold @xmath2 was introduced by scharlemann and thompson in @xcite .
the idea is to build the 3-manifold by starting with a set of 0-handles , then alternate between attaching collections of 1-handles and 2-handles keeping the boundary at the intermediate steps as simple as possible and finally add 3-handles .
such a decomposition of a manifold is called a generalized heegaard splitting .
the classical heegaard splitting where all 1-handles are attached at the same time followed by all 2-handles is an example of a generalized heegaard splitting .
casson and gordon @xcite show that if @xmath3 is a weakly reducible heegaard splitting for @xmath2 ( i.e. there are meridional disks for @xmath4 and @xmath5 with disjoint boundaries ) , then either @xmath3 is reducible or @xmath2 contains an essential surface .
scharlemann and thompson @xcite show that such surfaces arise naturally when a heegaard splitting in put in thin position .
suppose a closed orientable 3-manifold @xmath6 contains a link @xmath1 , then we can isotope @xmath1 so that it intersects each handlebody in boundary parallel arcs . in this case
we say that @xmath7 is a bridge surface for @xmath1 or that @xmath7 is a heegaard surface for the pair @xmath8 .
the idea was first introduced by schubert in the case that @xmath9 and @xmath10 and was extended by morimoto and sakuma for other 3-manifolds . in @xcite hayashi and shimokawa
considered multiple heegaard splittings for @xmath8 using the idea of changing the order in which the 1-handles and the 2-handles are attached .
they generalized the result of @xcite in this context , i.e. they showed that if @xmath7 is a strongly compressible bridge surface for @xmath1 , then either @xmath3 is stabilized or cancellable or @xmath11 contains an essential meridional surface .
in this paper we will generalize this important result one step further by weakening the hypothesis .
suppose @xmath2 is a compact orientable manifold and @xmath12 is a properly embedded surface transverse to a 1-submanifold @xmath13 . in some contexts
it is necessary to consider not only compressing disks for @xmath14 but also cut - disks , that is , disks whose boundary is essential on @xmath15 and that intersect @xmath16 in exactly one point , as for example in @xcite , @xcite and @xcite .
a bridge surface @xmath7 for a link @xmath1 is c - strongly compressible if there is a pair of disjoint cut or compressing disks for @xmath17 on opposite sides of @xmath7 . in particular
every strongly compressible bridge surface is c - strongly compressible .
we will show that if a bridge surface @xmath7 for @xmath1 is c - strongly compressible then either it can be simplified in one of four geometrically obvious ways or @xmath8 contains an essential meridional surface .
let @xmath2 be a compact orientable irreducible 3-manifold and let @xmath16 be a 1-manifold properly embedded in @xmath2 . a regular neighborhood of @xmath16 will be denoted @xmath18 . if @xmath19 is any subset of @xmath2 we will use @xmath20 to denote @xmath21 .
we will assume that any sphere in @xmath2 intersects @xmath16 in an even number of points .
as all the results we will develop are used in the context when @xmath16 only has closed components , this is a natural assumption . if @xmath1 is a link in @xmath2 , then any sphere in @xmath2 intersects @xmath1 in an even number of points , since the ball in bounds in @xmath2 contains no endpoints of @xmath1 .
suppose @xmath14 is a properly embedded surface in @xmath2 .
an _ essential curve _ on @xmath22 is a curve that does nt bound a disk on @xmath22 and it is not parallel to a puncture of @xmath22 . a _ compressing disk _
@xmath23 for @xmath22 is an embedded disk in @xmath24 so that @xmath25 is an essential curve on @xmath22 .
cut - disk _ is a disk @xmath26 such that @xmath27 is an essential curve on @xmath22 and @xmath28 .
c - disk _ is a cut or a compressing disk .
@xmath14 will be called _ incompressible _ if it has no compressing disks and _ c - incompressible _ if it has no c - disks .
@xmath14 will be called _ essential _ if it does not have compressing disks ( it may have cut disks ) , it is not boundary parallel in @xmath29 and it is not a sphere that bounds a ball disjoint from @xmath16 .
suppose @xmath30 is a compression body ( @xmath31 may have some sphere components ) .
a set of arcs @xmath32 is _ trivial _ if there is a homeomorphism after which each arc is either vertical , ie , @xmath33 or there is an embedded disk @xmath34 such that @xmath35 where @xmath36 . in the second case
we say that @xmath37 is @xmath38-parallel and the disk @xmath34 is a bridge disk . if @xmath30 is a handlebody , then all trivial arcs are @xmath38-parallel and are called _
bridges_. if @xmath16 is a 1-manifold properly embedded in a compression body @xmath30 so that @xmath16 is a collection of trivial arcs then we will denote the pair by @xmath39 .
let @xmath40 be a pair of a compression body and a 1-manifold and let @xmath41 be the disjoint union of compressing disks for @xmath42 together with one bridge disk for each @xmath38-parallel arc . if @xmath41 cuts @xmath40 into a manifold homeomorphic to @xmath43 _ vertical arcs_@xmath44 together with some 3-balls , then @xmath41 is called a _
complete disk system _ for @xmath40 .
the presence of such a complete disk system can be taken as the definition of @xmath40 .
let @xmath2 be a 3-manifold , let @xmath45 be a heegaard splitting ( ie @xmath4 and @xmath5 are compression bodies ) for @xmath2 and let @xmath16 be a 1-manifold in @xmath2 .
we say that @xmath16 is in bridge position with respect to @xmath7 if @xmath4 and @xmath5 intersect @xmath16 only in trivial arcs .
in this case we say that @xmath7 is a bridge surface for @xmath16 or that @xmath7 as a heegaard surface for the pair @xmath46 .
suppose @xmath47 and @xmath16 is in bridge position with respect to @xmath7 .
the heegaard splitting is _ c - strongly irreducible _ if any pair of c - disks on opposite sides of @xmath48 intersect , in this case the bridge surface @xmath48 is _ c - weakly incompressible_. if there are c - disks @xmath49 and @xmath50 such that @xmath51 , the heegaard splitting is _ c - weakly reducible _ and the bridge surface @xmath48 is _ c - strongly compressible_. following @xcite , the bridge surface @xmath48 will be called _ stabilized _ if there is a pair of compressing disks on opposite sides of @xmath48 that intersect in a single point .
the bridge surface is _ meridionally stabilized _ if there is a cut disk and a compressing disk on opposite sides of @xmath48 that intersect in a single point .
finally the bridge surface is called _ cancellable _ if there is a pair of canceling disks @xmath34 for bridges @xmath37 on opposite sides of @xmath7 such that @xmath52 .
if @xmath53 we will call the bridge surface _ perturbed_. in @xcite the authors show that is @xmath54 is stabilized , meridionally stabilized or perturbed , then there is a simpler bridge surface @xmath55 for @xmath16 such that @xmath7 can be obtained from @xmath55 by one of three obvious geometric operations .
if the bridge surface @xmath7 for @xmath16 is cancellable with canceling disks @xmath56 and @xmath57 such that @xmath58 then using this pair of disks some closed component @xmath59 of @xmath16 can be isotoped to lie in @xmath7 .
if this component can be isotoped to lie in the core of one of the compression bodies , @xmath4 say , and is disjoint from all other bridge disks in @xmath4 then @xmath60 is also a compression body and the 1-manifold @xmath61 intersects it in a collection of trivial arcs .
thus @xmath62 is heegaard splitting for @xmath63 and @xmath7 is a bridge surface for @xmath61 . in this case
we will say that @xmath16 has a _
removable component_. a detailed discussion of links with removable components is given in @xcite . in the absence of a knot
, it follows by a theorem of waldhausen that a heegaard splitting of an irreducible manifold is stabilized if and only if there is a sphere that intersects the heegaard surface in a single essential curve ( i.e the heegaard splitting is reducible ) , unless the heegaard splitting is the standard genus 1 heegaard splitting of @xmath64 .
we will say that a bridge surface for @xmath16 is _ c - reducible _ if there is a sphere or a twice punctured sphere in @xmath2 that intersects the bridge surface in a single essential closed curve .
then one direction of waldhausen s result easily generalizes to bridge surfaces as the next theorem shows .
[ thm : reducible ] suppose @xmath7 is a bridge surface for a 1-manifold @xmath16 properly embedded in a compact , orientable 3-manifold @xmath2 where @xmath7 is not the standard genus 1 heegaard splitting for @xmath64 . if @xmath7 is stabilized , perturbed or meridionally stabilized then there exists a sphere @xmath65 , possibly punctured by @xmath16 twice , which intersects @xmath7 in a single essential curve @xmath66 and neither component of @xmath67 is parallel to @xmath7 .
if @xmath7 is stabilized let @xmath65 be the boundary of a regular neighborhood of the union of the pair of stabilizing disks , figure [ fig : stabilizations ] . in this case
@xmath65 is a sphere disjoint from @xmath16 .
if @xmath7 is meridionally stabilized , let @xmath65 be the boundary of a regular neighborhood of the union of the cut and compressing disks . in this case
@xmath65 is a twice punctured sphere with both punctures on the same side of @xmath68 . finally if @xmath7 is perturbed with canceling disks @xmath69 and @xmath70 , let @xmath65 be the boundary of a regular neighborhood of @xmath71 . then @xmath65 is a twice punctured sphere and the punctures are separated by @xmath68 .
we will need to generalize the notion of a compression body containing trivial arcs as follows .
a c - compression body @xmath72 is a pair of a compression body @xmath30 and a 1-manifold @xmath16 such that there is a collection of disjoint bridge disks and c - disks @xmath73 so that @xmath73 cuts @xmath74 into a 3-manifold homeomorphic to @xmath75 vertical arcs@xmath44 together with some 3-balls . in this case
@xmath73 is called a complete c - disk system . one way to construct a compression body
is to take a product neighborhood @xmath76 of a closed , possibly disconnected , surface @xmath14 so that any arc of @xmath77 can either be isotoped to be vertical with respect to the product structure or is parallel to an arc in @xmath78 and then attach a collection of pairwise disjoint 2-handles @xmath79 to @xmath80 . if we allow some of the 2-handles in @xmath79 to contain an arc @xmath81 as their cocore , the resulting 3-manifold is a c - compression body .
the complete c - disk system described in the definition above consists of all bridge disks together with the cores of the 2-handles .
we will use this construction as an alternative definition of a c - compression body . _
recall that a spine of a compression body @xmath30 is the union of @xmath82 together with a 1-dimensional graph @xmath83 such that @xmath30 retracts to @xmath84 .
an equivalent definition of a c - compression body is that @xmath74 is a compression body @xmath30 together with a 1-manifold @xmath16 and there exists a spine @xmath85 for @xmath30 such that all arcs of @xmath16 that are not trivial in @xmath30 can be simultaneously isotoped to lie on @xmath85 and be pairwise disjoint .
we will however not use this definition here . _
[ prop : c - compcanbecomp ] let @xmath74 be a c - compression body .
then @xmath74 is a compression body if and only if there is no arc @xmath81 such that @xmath86 . in particular
if @xmath87 , then @xmath30 is a handlebody .
consider the construction above and note that before the two handles are added no arc of @xmath16 has both of its endpoints on @xmath80 .
if some 2-handle @xmath23 attached to @xmath80 contains an arc @xmath81 as its core , this arc will have both of its endpoints on @xmath88 .
thus @xmath30 is a compression body if and only if no 2-handle contains such an arc .
[ lem : surfaceincbody ] let @xmath74 be a c - compression body and let @xmath14 be a c - incompressible , @xmath89-incompressible properly embedded surface transverse to @xmath16 . then there is a complete c - disk system @xmath73 of @xmath74 such that @xmath90 consists of two types of arcs * an intersection arc @xmath66 between a bridge disk in @xmath73 and a twice punctured sphere component of @xmath14 with both endpoints of @xmath66 lying on @xmath16 . * an intersection arc @xmath91 between a bridge disk in @xmath73 and a once - punctured disk component of @xmath14 with one endpoint of @xmath91 lying on @xmath16 and the other lying on @xmath92 .
the argument is similar to the the proof of lemma 2.2 in @xcite so we only give an outline here .
let @xmath73 be a complete c - disk system for @xmath74 chosen to minimize @xmath93 .
using the fact that @xmath22 is c - incompressible , we may assume that @xmath90 does not contain any simple closed curves .
if @xmath94 is an arc with both of its endpoints on @xmath95 , then an outermost such arc either gives a @xmath89-compression for @xmath14 contrary to the hypothesis or can be removed by an outermost arc argument contradicting the minimality of @xmath93 . note that if @xmath66 lies on some cut - disk @xmath96 , we can still choose the arc so that the disk it cuts from @xmath96 does not contain a puncture .
this establishes that @xmath14 is disjoint from all c - disks in @xmath73 .
suppose @xmath66 is an arc of intersection between a bridge disk @xmath97 for @xmath16 and a component @xmath98 of @xmath14 .
assume that @xmath66 is an outermost such arc and let @xmath99 be the subdisk @xmath66 bounds on @xmath97 . by the above argument at least one endpoint of @xmath66 must lie on @xmath16 .
if both endpoints of @xmath66 lie on @xmath16 , the boundary of a regular neighborhood of @xmath100 gives a compressing disk for @xmath14 contrary to the hypothesis unless @xmath98 is a twice punctured sphere . if @xmath66 has one endpoint on @xmath16 and one endpoint on @xmath95 , a regular neighborhood of @xmath100 is a @xmath89-compressing disk for @xmath14 unless @xmath98 is a once punctured disk .
[ cor : negativeboundaryincomp ] if @xmath74 is a c - compression body , then @xmath88 is incompressible .
suppose @xmath23 is a compressing disk for some component of @xmath88 .
by lemma [ lem : surfaceincbody ] , there exists a complete c - disk system @xmath73 for @xmath74 such that @xmath101 . but
this implies that @xmath23 is a @xmath89-reducing disk for the manifold @xmath102 _ vertical arcs_@xmath44 , a contradiction . if @xmath2 is a 3-manifold we will denote by @xmath103 the manifold obtained from @xmath2 by filling any sphere boundary components of @xmath2 with 3-balls .
[ lemma 2.4 @xcite][lem : surfaceinproduct ] if @xmath14 is an incompressible , @xmath89-incompressible surface in a compression body @xmath39 , then @xmath14 is a collection of the following kinds of components : * spheres intersecting @xmath16 in 0 or 2 points , * disks intersecting @xmath16 in 0 or 1 points , * vertical annuli disjoint from @xmath16 , * closed surfaces parallel to a component of @xmath104 .
[ cor : components ] if @xmath14 is a c - incompressible , @xmath89-incompressible surface in a c - compression body @xmath74 , then @xmath14 is a collection of the following kinds of components : * spheres intersecting @xmath16 in 0 or 2 points , * disks intersecting @xmath16 in 0 or 1 points , * vertical annuli disjoint from @xmath16 , * closed surfaces parallel to a component of @xmath104 . delete all component of the first two types and let @xmath98 be the new surface . by lemma [ lem : surfaceincbody ] , there exists a complete c - disk system @xmath41 for @xmath74 such that @xmath105 .
thus each component of @xmath98 is contained in a compression body with trivial arcs ( in fact in a trivial compression body but we do nt need this fact ) .
the result follows by lemma [ lem : surfaceinproduct ] .
the following definition was first introduced in @xcite if @xmath16 is a 1-manifold properly embedded in a compact 3-manifold @xmath2 , we say that the disjoint union of surfaces @xmath106 is a multiple heegaard splitting of @xmath46 if 1 .
the closures of all components of @xmath107 are compression bodies @xmath108 , 2 . for @xmath109 , @xmath110 is attached to some @xmath111 where @xmath112 , 3 .
a component of @xmath113 is attached to some component of @xmath114 ( possibly @xmath115 ) .
a component @xmath0 of @xmath106 is said to be positive if @xmath116 for some @xmath117 and negative if @xmath118 for some @xmath119 .
the unions of all positive and all negative components of @xmath106 are denoted @xmath120 and @xmath121 respectively .
note that if @xmath106 has a single surface component @xmath7 , then @xmath7 is a bridge surface for @xmath16 .
using c - compression bodies instead of compression bodies , we generalize this definition as follows .
if @xmath16 is a 1-manifold properly embedded in a compact 3-manifold @xmath2 , we say that the disjoint union of surfaces @xmath106 is a multiple c - heegaard splitting of @xmath46 if 1 .
the closures of all components of @xmath107 are c - compression bodies @xmath122 , 2 . for @xmath109 , @xmath110 is attached to some @xmath111 where @xmath112 , 3 .
a component of @xmath113 is attached to some component of @xmath114 ( possibly @xmath115 ) as in @xcite and @xcite we will associate to a multiple c - heegaard splitting a measure of its complexity . the following notion of complexity of a surface is different from the one used in @xcite .
let @xmath65 be a closed connected surfaces embedded in @xmath2 transverse to a properly embedded 1-manifold @xmath123 .
the complexity of @xmath65 is the ordered pair @xmath124 . if @xmath65 is not connected , @xmath125 is the multi - set of ordered pairs corresponding to each of the components of @xmath65 . as in @xcite the complexities of two possibly not connected surfaces
are compared by first arranging the ordered pairs in each multi - set in non - increasing order and then comparing the two multi - sets lexicographically where the ordered pairs are also compared lexicographically .
[ lem : compreduces ] suppose @xmath126 is meridional surface in @xmath46 of non - positive euler characteristic . if @xmath127 is a component of the surface obtained from @xmath126 by compressing along a c - disk , then @xmath128 . without loss of generality
we may assume that @xmath126 is connected .
_ case 1 : _ let @xmath129 be a possibly disconnected surface obtained from @xmath126 via a compression along a disk @xmath23 . in this case
@xmath130 as @xmath131 so the result follows immediately if @xmath129 is connected .
if @xmath129 consists of two components then by the definition of compressing disk , we may assume that neither component is a sphere and thus both components of @xmath129 have non - positive euler characteristic .
by the additivity of euler characteristic it follows that if @xmath132 is a component of @xmath129 , then @xmath133 so @xmath134 as desired . _ case 2 : _
suppose @xmath129 is obtained from @xmath126 via a compression along a cut - disk @xmath96 .
if @xmath96 is separating , then each of the two components of @xmath129 has at least one puncture and if a component is a sphere , then it must have at least 3 punctures , ie each component of @xmath129 has a strictly negative euler characteristic . by the additivity of euler characteristic ,
we conclude that for each component @xmath132 of @xmath135 , @xmath136 and so the first component of the complexity tuple is decreased .
if the cut disk is not separating the cut - compression does not affect the first term in the complexity tuple as @xmath137 . note that @xmath138 must be essential in the non - punctured surface @xmath65 so we can consider @xmath96 as a compressing disk for @xmath65 in @xmath2 .
then @xmath139 so in this case the second component of the complexity tuple is decreased .
the _ width _ of a c - heegaard splitting is the multiset of pairs @xmath140 . in @xcite
a multiple heegaard splitting is called _ thin _ if it is of minimum width amongst all possible multiple heegaard splittings for the pair @xmath46 .
similarly we will call a c - heegaard splitting _
c - thin _ if it is of minimal width amongst all c - heegaard splittings for @xmath46 .
[ lem : thinning ] suppose @xmath2 is a compact orientable irreducible manifold and @xmath16 is a properly embedded 1-submanifold .
if @xmath7 is a c - heegaard splitting for @xmath46 which is c - weakly reducible , then there exists a multiple c - heegaard splitting @xmath141 so that @xmath142 .
moreover if @xmath2 is closed then either * there is a component of @xmath143 that is neither an inessential sphere nor boundary parallel in @xmath24 , or * @xmath7 is stabilized , meridionally stabilized or perturbed , or a closed component of @xmath16 is removable .
the first part of the proof of this lemma is very similar to the proof of lemma 2.3 in @xcite and uses the idea of _
untelescoping_. however , in lemma 2.3 the authors only allow untelescoping using disks while we also allow untelescoping using cut - disks .
let @xmath144 and @xmath145 be the two c - compression bodies that @xmath7 cuts @xmath146 into . consider a maximal collection of c - disks @xmath147 and @xmath148 such that @xmath149 .
let @xmath150 and @xmath151 where @xmath152 is a collar of @xmath153 .
then @xmath154 and @xmath155 are each the disjoint union of c - compression bodies .
take a small collar @xmath156 of @xmath157 and @xmath158 of @xmath159 .
let @xmath160 , @xmath161 , @xmath162 and @xmath163 .
this is a new multiple c - heegaard splitting of @xmath46 with positive surfaces @xmath164 and @xmath165 that can be obtained from @xmath7 by c - compressing along @xmath166 and @xmath167 respectively and a negative surface @xmath168 obtained from @xmath7 by compressing along both sets of c - disks . by lemma [ lem : compreduces ]
it follows that @xmath169 . to show the second part of the lemma , suppose @xmath3 is not stabilized , meridionally stabilized or perturbed and no component of @xmath16 is removable and , by way of contradiction ,
suppose that every component of @xmath170 is a sphere bounding a ball that intersects @xmath16 in at most one trivial arc or a torus that bounds a solid torus such that @xmath81 is a core curve of it .
let @xmath171 and @xmath172 be the arcs that are the cocores of the collections of c - disks @xmath166 and @xmath167 respectively . if @xmath96 is a cut - disk , we take @xmath173 as its cocore .
let @xmath174 and note that @xmath7 can be recovered from @xmath175 by surgery along @xmath176 .
as @xmath7 is connected , at least one component of @xmath175 must be adjacent to both @xmath171 and @xmath172 , call this component @xmath14 . unless @xmath22 is is an inessential sphere or boundary parallel in @xmath24 we are done . if @xmath22 is an inessential sphere , then by waldhausen s result the original heegaard splitting is stabilized . as @xmath177 by hypothesis ,
the remaining possibility is that @xmath22 is parallel in @xmath24 to part of @xmath16 ; since @xmath22 is connected it is either a torus bounding a solid torus with a component of @xmath16 as its core or @xmath22 is an annulus , parallel to a subarc of @xmath16 .
that is @xmath14 bounds a ball which @xmath16 intersects in a trivial arc .
let @xmath178 be the ball or solid torus @xmath14 bounds .
we will assume that @xmath178 lies on the side of @xmath14 that is adjacent to @xmath171 and that @xmath14 is innermost in the sense that @xmath179 .
let @xmath180 and let @xmath181 be the c - compression body obtained by adding the 1-handles corresponding to the arcs @xmath182 to a collar of @xmath183 .
( some of these 1-handles might have subarcs of @xmath16 as their core ) . let @xmath184 .
notice that @xmath185 can be obtained from @xmath5 by c - compressing along all c - disks whose cocores are adjacent to @xmath14 and thus @xmath185 is a c - compression body .
in fact @xmath186 so @xmath185 is a handlebody , let @xmath187 . then @xmath188 is a c - heegaard splitting for @xmath178 decomposing in into a c - compression body @xmath181 and a handlebody @xmath185 .
there are two cases to consider : @xmath178 being a ball intersecting @xmath16 is a trivial arc and @xmath178 being a torus
. we will consider each case separately and prove that @xmath188 is actually a heegaard splitting for @xmath178 ( i.e. @xmath181 is a compression body ) so we can apply previously known results .
* case 1 : * if @xmath178 is a ball and @xmath189 is a trivial arc , there are three sub - cases to consider . if @xmath190 then the construction above gives a nontrivial heegaard splitting for the pair @xmath191 ; @xmath181 is a compression body by proposition [ prop : c - compcanbecomp ] as @xmath192 adjacent to two subarcs of @xmath59 both of which have their second endpoint on @xmath193 . by lemma 2.1 of @xcite
, @xmath194 is either stabilized or perturbed ( in this context if @xmath194 is cancellable , it must be perturbed as @xmath59 is not closed ) so the same is true for @xmath7 . if @xmath195 and @xmath196 ( in particular @xmath197 ) , figure [ fig : merstab ] shows a pair of c - disks demonstrating that @xmath7 is meridionally stabilized . if @xmath195 and @xmath198 , consider the solid torus @xmath199
let @xmath200 be the c - compression body obtained by 1-surgery on @xmath0 along the arcs @xmath201 .
as @xmath202 , @xmath200 is in fact a compression body . note that @xmath203 as @xmath204 . thus @xmath205 is a non - trivial heegaard splitting for the solid torus @xmath206 . by @xcite it
must be stabilized and thus so is @xmath7 . *
case 2 : * suppose @xmath14 bounds a solid torus @xmath178 , which is a regular neighborhood of closed component @xmath59 of @xmath16 .
as @xmath207 , @xmath188 is a heegaard splitting for @xmath208 .
by @xcite it is cancellable or stabilized .
this proves the theorem at hand unless @xmath194 is cancellable but not perturbed so assume this is the case .
in particular this implies that @xmath209 . in this case @xcite
shows that if @xmath210 then @xmath194 is stabilized .
thus it remains to consider the case when @xmath194 is a torus intersecting @xmath59 in two points .
in this case @xmath0 must be the union of @xmath14 and a sphere @xmath65 intersecting @xmath59 in two points and @xmath211 is a single possibly knotted arc with one endpoint on @xmath14 and the other on @xmath65 .
as @xmath59 is cancellable , we can use the canceling disk in @xmath181 to isotope @xmath59 across @xmath194 so it lies entirely in @xmath185 .
after this isotopy it is clear that @xmath14 and @xmath194 cobound a product region .
as @xmath14 is the boundary of a regular neighborhood of @xmath59 , it follows that @xmath59 is isotopic to the core loop of the solid torus @xmath185 ie , @xmath212 is a trivial compression body .
@xmath5 can be recovered from @xmath185 by 1-surgery so @xmath213 is also a compression body .
thus after an isotopy of @xmath59 along the pair of canceling disks , @xmath7 is a heegaard splitting for @xmath214 so @xmath59 is a removable component of @xmath16 .
as in jaco @xcite a weak hierarchy for a compact orientable 2-manifold @xmath14 is a sequence of pairs @xmath215 where @xmath216 , @xmath217 is an essential curve on @xmath218 and @xmath219 is obtained from @xmath218 by cutting @xmath218 along @xmath217 .
the final surface in the hierarchy , @xmath220 , satisfies the following : 1 .
each component of @xmath220 is a disc or an annulus at least one boundary component of which is a component of @xmath221 .
each non - annulus component of @xmath14 has at least one boundary component which survives in @xmath222 .
the following lemma was first proven by jaco and then extended in @xcite , lemma 3.1 .
[ lem : boundarynumber ] let @xmath14 be a connected planar surface with @xmath223 boundary components .
let @xmath224 be a weak hierarchy with each @xmath217 an arc .
if @xmath225 is the number of boundary components of @xmath220 then , * if @xmath220 does not have annulus components then @xmath226 * if @xmath220 contains an annulus component , then @xmath227 . if @xmath228 and @xmath229 , then @xmath220 contains a disc component .
[ thm : one ] suppose @xmath2 is a compact orientable irreducible manifold and @xmath16 is a properly embedded 1-manifold in @xmath2 .
let @xmath45 be a c - heegaard splitting for @xmath46 .
if @xmath23 is a boundary reducing disk for @xmath2 then there exists such disk @xmath230 so that @xmath230 intersects @xmath48 in a unique essential simple closed curve .
let @xmath23 be a reducing disk for @xmath231 chosen amongst all such disks so that @xmath232 is minimal .
by corollary [ cor : negativeboundaryincomp ] , @xmath233 .
let @xmath234 and @xmath235 .
suppose some component of @xmath236 is c - compressible in @xmath4 with @xmath97 the c - compressing disk .
let @xmath237 and let @xmath238 be the disk @xmath239 bounds on @xmath23 .
note that the sphere @xmath240 must be punctured by @xmath16 either 0 or two times thus @xmath97 must be a non - punctured disk .
let @xmath241 .
@xmath230 is also a reducing disk for @xmath231 as @xmath242 and @xmath243 . as @xmath244 is essential on @xmath236 , @xmath245
can not lie entirely in @xmath4 so @xmath246 and thus @xmath247 contradicting the choice of @xmath23 . similarly @xmath248 is c - incompressible in @xmath5 .
suppose that @xmath97 is a @xmath89-compressing disk for @xmath236 and @xmath97 is adjacent to @xmath249 .
@xmath89-compressing @xmath23 along @xmath97 gives two disks @xmath56 and @xmath57 at least one of which has boundary essential of @xmath231 , say @xmath56 .
however @xmath250 , a contradiction .
suppose that @xmath97 is a @xmath89-compressing disk for @xmath236 and @xmath97 is adjacent to @xmath7 .
let @xmath251 .
use @xmath97 to isotope @xmath23 so that a neighborhood of @xmath66 lies in @xmath5 , call this new disk @xmath252 and let @xmath253 and @xmath254 . note that @xmath255 is obtained from @xmath236 by cutting along
@xmath66 and @xmath255 is also c - incompressible .
repeat the above operation naming each successive disk @xmath256 until the resulting surface @xmath257 is @xmath89-incompressible . by corollary [ cor : components ] @xmath258
consists of vertical annuli and disks .
suppose some component of @xmath236 is @xmath89-compressible but not adjacent to @xmath259 .
in this case the result of maximally @xmath89-compressing this component has to be a collection of disks . by case 1 of lemma [ lem : boundarynumber ] ,
@xmath260 contradicting our choice of @xmath23 .
thus every boundary compressible component of @xmath236 is adjacent to @xmath249 , in particular @xmath261 and @xmath236 has a unique @xmath89-compressible component @xmath14 . by the minimality assumption and case 2 of lemma [ lem : boundarynumber ] , some component of @xmath258 must be a disk .
@xmath262 is then a planar surface that we have shown must be c - incompressible and has a component that is not a disk . as @xmath263
, it follows that some component of @xmath262 is @xmath89-compressible into @xmath7 and disjoint from @xmath264 .
the above argument applied to @xmath262 leads to an isotopy of the disk @xmath23 so as to reduce @xmath265 contrary to our assumption .
thus @xmath236 and @xmath248 are both collections of vertical annuli and disks so @xmath23 is a reducing disk for @xmath231 that intersects @xmath7 in a single essential simple closed curve .
[ cor : stronglyirreducible ] let @xmath3 be a c - strongly irreducible c - heegaard splitting of @xmath46 and let @xmath14 be a component of @xmath231 .
if @xmath22 is not parallel to @xmath48 , then @xmath22 is incompressible .
suppose @xmath23 is a compressing disk for @xmath266 say . by theorem [ thm : one ]
we can take @xmath23 such that @xmath267 , @xmath268 is a compressing disk for @xmath48 lying in @xmath4 and @xmath269 is a vertical annulus disjoint from @xmath16 . as @xmath22 is not parallel to @xmath48
, there is a c - disk for @xmath48 lying in @xmath4 , @xmath236 . by a standard innermost disk and outermost arc arguments
, we can take @xmath236 so that @xmath270 .
but then @xmath236 and @xmath248 give a pair of c - weakly reducing disks for @xmath48 contrary to our hypothesis .
following @xcite we will call a c - heegaard splitting @xmath106 _ c - slim _ if each component @xmath271 obtained by cutting @xmath2 along @xmath121 is c - strongly irreducible and no proper subset of @xmath106 is also a multiple c - heegaard splitting for @xmath2 .
suppose @xmath106 is a c - thin c - heegaard splitting of @xmath2 . if some proper subset of @xmath106 is also a c - heegaard splitting of @xmath2 , then this c - heegaard splitting will have lower width than @xmath106 . if some component @xmath272 of @xmath107 is c - weakly reducible , applying the untelescoping operation described in lemma [ lem : thinning ] to that component produces a c - heegaard splitting of lower width . thus if @xmath106 is c - thin , then it is also c - slim .
[ thm : essentialexists ] suppose @xmath2 is a closed orientable irreducible 3-manifold containing a link @xmath1 . if @xmath7 is a c - strongly compressible bridge surface for @xmath1 then one of the following is satisfied : * @xmath7 is stabilized * @xmath7 is meridionally stabilized * @xmath7 is perturbed * a component of @xmath1 is removable * @xmath2 contains an essential meridional surface @xmath14 such that @xmath273 .
let @xmath106 be a c - slim heegaard splitting obtained from @xmath7 by untelescoping as in lemma [ lem : thinning ] , possibly in several steps .
let @xmath121 and @xmath120 denote the negative and positive surfaces of @xmath106 respectively and let @xmath272 be the components of @xmath274 where @xmath272 is the union of c - compression bodies @xmath275 and @xmath276 along @xmath277 .
suppose some component of @xmath121 is compressible with compressing disk @xmath23 . by taking an innermost on @xmath23 circle of @xmath278
we may assume that @xmath113 is compressible in @xmath272 .
by corollary [ cor : stronglyirreducible ] this contradicts our assumption that @xmath106 is c - slim .
we conclude that @xmath121 is incompressible . if some component of @xmath279 of @xmath121 is neither an inessential sphere nor boundary parallel in @xmath280 , then it is essential and @xmath273 . if every component is either an inessential sphere in @xmath280 or boundary parallel , then by lemma [ lem : thinning ] the splitting is perturbed , stabilized , meridionally stabilized or there is a removable component
i would like to thank martin scharlemann for many helpful conversations .
martin scharlemann and abigail thompson .
thin position for @xmath282-manifolds . in
_ geometric topology ( haifa , 1992 ) _ , volume 164 of _ contemp .
_ , pages 231238 .
soc . , providence , ri , 1994 . | we extend the notion of thin multiple heegaard splittings of a link in a 3-manifold to take into consideration not only compressing disks but also cut - disks for the heegaard surfaces . we prove that if @xmath0 is a c - strongly compressible bridge surface for a link @xmath1 contained in a closed orientable irreducible 3-manifold @xmath2 then one of the following is satisfied : * @xmath0 is stabilized * @xmath0 is meridionally stabilized * @xmath0 is perturbed * a component of @xmath1 is removable * @xmath2 contains an essential meridional surface . |
tau leptons are an excellent signature with which to probe new physics at the lhc . as the heaviest leptons
, they have the largest coupling to the higgs boson both in the standard model ( sm ) and the minimal supersymmetric standard model ( mssm ) . their short - enough lifetime and parity - violating decays allow for their spin information to be preserved in the decay product kinematics recorded in the detector .
they are therefore the only leptons suitable for measuring longitudinal polarization and its correlations , which provide important constraints on the nature of the observed resonance . in this note
we present a tauspinner package , which is a mc program designed to generate tau spin effects in _ any _ tau sample provided their origin is known .
this tool has two important applications : data driven analysis : : of @xmath3 and @xmath4 + backgrounds . in the case of the algorithms for embedding taus on
measured light lepton samples @xcite the tauspinner may represent a third step .
the first step is to construct the tau four - momenta from the four - momenta of the measured lighter leptons ( and accompanying photons ) while the second step comprises the decay of unpolarized taus using + e.g. tauola @xcite .
omission of the last step leaves the kinematics of the tau decay particle different from what is expected from the polarized taus that appear in nature .
this could lead to a mis - modeling of the shapes of various observables and hence a mis - measurement of the acceptance of a given set of cuts .
mc studies .
: : since the tau spin effects affect the overall acceptance of the taus in the detector , a proper treatment of these effects is of great importance for the interpretation of the results as well as feasibility studies of new models .
it is also key for measurements of tau polarization and tau spin correlations .
the tauspinner algorithm allows one to create samples with different tau polarizations from an initial sample by re - weighting events to give the desired distributions as a function of the decay mode of the tau .
furthermore , the helicity states of the taus can be attributed to the previously generated sample ( with spin effect included or introduced with the tauspinner weight ) .
this letter is organized as follows . in sec .
[ sect : thealg ] we present the algorithm while in sec .
[ sect : perf ] we discuss its performance as compared to the standard tauola mc package .
the results are summarized in sec .
[ sect : summary ] .
all relevant technical details are provided in app .
[ sect : thealg ] the tauspinner algorithm relies on a leading order approximation in which spin amplitudes are used to calculate the spin density matrices for hard @xmath5 or @xmath6 born level processes @xcite .
tauspinner is constrained to the longitudinal tau spin degrees only .
it starts with identifying the flavor of the intermediate boson : @xmath7 , @xmath2 , @xmath8 or @xmath9 .
the information on the four - momenta of the outgoing taus and their decay products as well as the intermediate boson four - momentum is then used to determine the polarimetric vectors . the longitudinal tau polarization ( @xmath10 )
is randomly generated as specified in tab .
[ tab : spinconf ] and set to @xmath11 which correspond to pure tau helicity states .
probability of the helicity states is a constant for taus originating from the @xmath7 , @xmath9 and @xmath8 bosons . in @xmath3 events , the probability , denoted @xmath12 ,
is a function of the @xmath13 scattering angle , @xmath14 , and the center of mass squared of the hard process , @xmath15 . at the born level , and in the ultrarelativistic limit , @xmath12 is , following notations of @xcite , given by : @xmath16 where @xmath17,\ ] ] and @xmath18 s are four form factors which depend on the initial and the final state fermion couplings to the @xmath19 boson and the propagator .
the dependence on the @xmath13 longitudinal polarization @xmath20 is taken into account ( note that in this case @xmath211 , also @xmath22 ) . in
@xcite the probability @xmath12 is calculated using the information on the initial state quarks stored at the generation level . in tauspinner
the initial state quark configuration is attributed stochastically from the intermediate boson kinematics and pdfs in the following steps : 1 .
the invariant mass of the @xmath2 is calculated from the intermediate boson four - vector .
note that it does not need to coincide with the sum of @xmath0 four - momenta as it may include photons of final state bremsstrahlung .
the scattering angle , @xmath23 , is calculated in the @xmath0 + pair rest frame from the angle between the direction of the first beam ( 1,0,0,1 ) boosted to this frame and the direction of the @xmath24 or from the angle between the direction of the second beam ( 1,0,0,-1 ) boosted to this frame and the direction of the @xmath13 . in the final step the average of the two is taken @xcite ] .
the fraction of momenta taken by partons of the proton : @xmath25 and @xmath26 are resolved from constraints + @xmath27 and @xmath28 , where @xmath29 + and @xmath30 denote the collision center of mass energy and the longitudinal component of the @xmath2 momentum , respectively .
probabilistic choice on the basis on the leading order @xmath31 born level cross sections and pdf s is performed to attribute the flavors to the incoming quarks and sign of @xmath14 .
the probability @xmath12 is calculated as a weighted average over all possible initial state quark configurations .
this solution is only implemented for the case of proton - proton ( @xmath32 ) collisions . for other types of collision events
the functionality of tauspinner for @xmath3process is restricted to longitudinal spin correlation only ( @xmath12=0.5 ) .
note that the calculation of the momentum fractions @xmath33 and @xmath34 is performed within a collinear approximation where the initial state radiation quarks and gluons are assumed to have no transverse momentum .
furthermore the final state radiation photons are taken into account at collinear level only : they are omitted from the boson decay vertex .
any deviation from the energy - momentum conservation in the boson decay vertex is attributed to the presence of the photons . as shown in sec .
[ sect : perf ] these approximations have practically no impact on the implemented spin effects .
lccc origin & @xmath35 & @xmath36 & probability + neutral higgs bosons : @xmath8 & + 1 & 1&0.5 + & 1 & + 1&0.5 + neutral vector boson : @xmath2 & + 1 & + 1&@xmath12 + & 1 & 1&1@xmath12 + charged higgs : @xmath9 & + 1 & & 1.0 + charged vector boson : @xmath7 & 1 & & 1.0 + the outcome of running the tauspinner program is a spin weight attributed to each event separately . in the @xmath1 final state , for any decay of a polarized tau , the spin weight is defined as @xcite : @xmath37 where @xmath38 is the tau polarization vector , and @xmath39 is the polarimetric vector constructed using hadronic currents . in our @xmath40 and @xmath41 decays the exact expression reduces to : @xmath42 where @xmath43 equals one for left - handed taus from @xmath7 bosons and minus one for right - handed taus from the charged higgs boson .
@xmath44 is the @xmath45 component of the polarimetric vector . in the @xmath0 final state ,
the weight is defined as @xcite : @xmath46 where @xmath47 is a matrix describing the full spin correlation between the two taus as well as the individual spin states of the taus . @xmath48 and @xmath49 are the time ( @xmath50 ) and space ( @xmath51 ) components of the two taus polarimetric vectors .
the @xmath50 component of @xmath52 and @xmath53 are by convention set to @xmath54 . neglecting the transverse spin degree , and in the ultrarelativistic limit , for @xmath3 events the expression reduces to : @xmath55 where @xmath10 denotes the polarization of the single tau in a mixed quantum state . within this approximation ,
@xmath10 is a linear function of the probability @xmath12 : @xmath56 in an event of a neutral and spin zero higgs boson decaying to @xmath0 , the expression simplifies to : @xmath57 the @xmath43 equals one for the @xmath2 boson and minus one for the neutral higgs boson , reflecting the opposite spin correlations in the two samples .
each tau decay channel requires a distinct method to calculate the polarimetric vector @xcite .
the tau decay modes implemented in tauspinner are listed in tab .
[ tab : taumodes ] . for the remaining channels , involving five pions in the final state and multi - prong decays with kaons ,
the effect of tau polarization is neglected and @xmath58 is set to zero .
lc tau decay mode & branching fraction % + @xmath59 & 17.85 + @xmath60 & 17.36 + @xmath61 & 10.91 + @xmath62 & 25.51 + @xmath63 , @xmath64 & 9.29 , 9.03 ( incl .
@xmath65 ) + @xmath66 & 0.70 + @xmath67 , @xmath68 & 0.43 , 0.84 + @xmath69 & 4.54 ( incl .
@xmath65 ) + @xmath70 & 1.04 + other & 2.5 + the event weight can be used at the analysis level for : simulating tau spin effects : : e.g. in a sample generated without spin effects .
the event weight equals @xmath71 .
it takes values between ( 0,2 ) , except for the case of @xmath3 when the range is ( 0,4 ) . removing tau spin effects
: : from a sample generated with spin effects .
the event weight equals 1@xmath72 .
then , it is greater than zero with no upper limit .
spin effects : : in a sample generated with certain longitudinal tau polarization ( and/or correlations ) to the different one .
the weight equals ( 2-@xmath71)/@xmath71 for the @xmath1 final state corresponding to the @xmath7@xmath73@xmath9 or @xmath9@xmath73@xmath7 replacement , and @xmath71(@xmath8)/@xmath71(@xmath2 ) or @xmath71(@xmath2)/@xmath71(@xmath8 ) for the @xmath0 + final state corresponding to @xmath2@xmath73@xmath8 or @xmath8@xmath73@xmath2 replacement .
it is greater than zero without an upper limit .
the helicity states of the taus are attributed stochastically by comparing a random number with the probability , @xmath12 @xmath74@xmath12@xmath75 , of the right - handed configuration to occur .
this method is valid for events generated with spin effects or emulated using the spin weight provided by the tauspinner .
an average over all possible initial state configuration is taken .
presentation of the method principle can be found in ref .
the performance of the tauspinner algorithm is studied based on the mc simulated @xmath3 and @xmath4 events .
the samples were generated using the general purpose event generator pythia @xcite assuming the @xmath32 collision at center of mass energy of 7 tev .
the taus were then made to decay using the tauola package .
two sets of events were simulated : _ no spin effects . _ : : in these events the taus were decayed by tauola as if they had been produced with no polarization .
they were used for emulation of spin effects later , with the tauspinner package .
_ : : in these events tau spin effects were properly accounted for at the time of tau event generation and decay . event generation and simulation of the spin effects in the tauspinner were performed with the modified lo parton distribution function ( pdf ) mrstlo * @xcite .
for all tau decay modes , the main observables that are sensitive to the tau polarization are the tau momentum fraction taken by the hadronic system @xcite , x , and the relative difference between the charge and the neutral energy in the tau decay , @xmath76 @xcite .
plots a)c ) in figure [ fig : taupolar ] demonstrates the performance of tauspinner for the channels where the tau polarization is extremal . the observable x is plotted in fig .
[ fig : taupolar1 ] for the combined @xmath77 and @xmath78 channels .
the observable @xmath76 is plotted in fig .
[ fig : taupolar2 ] and [ fig : taupolar3 ] for the @xmath79 and combined @xmath80 and @xmath81 channels , respectively .
the sample with no spin effects refers to @xmath4 + events generated with flat @xmath10 value .
the @xmath82 and @xmath4 configurations were obtained by applying an appropriate spin weight to the sample with no spin effects .
the weighted observables exhibit the expected behavior indicating a proper implementation of the @xmath10 in the tauspinner package .
note that the tauspinner algorithm requires that the decays of unpolarized taus via @xmath83 or @xmath84 preserve the spin correlation between the production and decay of these mesons .
this feature is missing if the taus are made to decay using pythia 6.425 .
although a bulk of spin effects are reconstructed , a significant systematic uncertainty arises as demonstrated in fig .
[ fig : taupolar ] d ) . in @xmath3 events ,
the single tau polarization @xmath10 depends on the intermediate boson virtuality , the flavor of the incoming quark and the scattering angle .
the complexity of this dependence is shown in fig .
[ fig : taupolcos ] where the tau polarization is drawn as a function of @xmath23 , for two different ranges of the invariant mass of the boson . in these plots ,
the quark flavor configuration is fixed at the level of simulation of the spin weight in the tauspinner . in the next step , a degree of polarization
is studied for different configurations of the incoming quarks .
the virtuality of the intermediate boson is constrained to lie within @xmath853 gev of the @xmath19 boson mass and both taus are set to decay to @xmath77 or @xmath78 .
the forward - backward spin asymmetry is accessed by choosing the @xmath13 to be emitted in the forward region by requiring the longitudinal momentum of @xmath13 to be greater than that of @xmath24 .
figure [ fig : xfricc ] shows the observable @xmath86 for the up and down type quarks entering along the positive or negative @xmath45 axis .
the results are consistent with those of reference @xcite , and therefore reassure a proper transmission of spin effects from the hard process to the tau decay products . in the last step
, the degree of polarization is studied inclusively for all initial state quark configurations and the results are compared to those simulated using the tauola package .
the @xmath2 virtuality is constrained to lie within @xmath853 gev of the @xmath19 boson mass .
the fraction of the tau momentum taken by the hadron in the @xmath77 channel is plotted in fig .
[ fig : xfrfb ] . expecting the @xmath19 boson to be emitted in the direction of the incoming quark ( as compared to the direction of the anti - quark from the sea ) these plots can be compared to those in fig .
[ fig : xfricc ] .
figure [ fig : tausplitting ] shows the fraction of the tau momentum taken by the hadron in the combined @xmath77 and @xmath78 channels for the left - handed and the right - handed taus from the @xmath3decays .
helicity states are attributed by the tauspinner in the sample without and with spin effects simulated at the generation level .
the distributions exhibit the expected shapes for properly attributed pure helicity states . and @xmath78 channels . @xmath87 and
@xmath88 denote right - handed and left - handed taus , respectively.,scaledwidth=45.0% ] two observables were demonstrated in @xcite to be appropriate for studying tau spin correlations in the @xmath0 final state : the invariant mass of the hadronic system , @xmath89 , and the z@xmath90 variable .
the latter is a signed surface in the x@xmath91 - x@xmath92 plane , between lines : x@xmath91=x@xmath92 and x@xmath91=x@xmath92+a . x@xmath91 and x@xmath92 denote the fraction of tau momenta taken by the hadrons .
the two observables are plotted in fig .
[ fig : taucorr ] in the channel where both taus decay to @xmath77 or @xmath78 .
the sample with no spin effects refers to @xmath3 events generated with @xmath12=0.5 and no spin correlation .
the proper @xmath93 and @xmath3 configurations were obtained by applying an appropriate spin weight to this sample without spin effects .
the weighted observables exhibit the expected behavior reassuring a proper implementation of the spin correlations in the tauspinner package .
the tauspinner package designed to emulate tau spin effects has been introduced .
the algorithm is limited to the leading order accuracy and the longitudinal spin degrees only . as compared to the algorithm developed in ref .
@xcite , the functionality is extended to allow one to estimate the spin effects when the information on the incoming quarks entering the hard process is not available .
this is therefore the first algorithm suitable for emulating the spin effects in the tau - embedded samples .
the intrinsic tau polarization arising from parity violation in the weak interactions is attributed on the basis of intermediate boson kinematics and pdf s . for the @xmath3 process
its complete functionality is limited to @xmath32 collisions and otherwise restricted to conservation of total angular momentum .
comparisons of tauola and tauspinner for various spin observables presented in sec .
[ sect : perf ] demonstrate that no additional systematic uncertainty in simulation of tau spin effects has been introduced in the tauspinner .
this claim should , however , be validated in a sample where the number of high transverse momentum jets is enhanced .
a complete discussion of the theoretical uncertainties is common for the tauola , the tauspinner and , to a large extent the koralz @xcite mc programs .
some aspects of these uncertainties have been addressed through the work published in refs .
@xcite . a rigorous evaluation of theoretical effects and in particular a comparison with results based on the next - to - leading order matrix element calculations simultaneous for scattering processes and for spin density matrices is referred to a future work .
although the current version of the tauspinner is restricted to the longitudinal spin degrees , the framework is prepared to simulate the complete spin effects .
this extension is planned for the future version of the algorithm .
the code of tauspinner is publicly available , with all relevant details given in app .
we would like to thank a. buckley , s.demers , c. gwenlan , e. richter - was and s. tsuno for valuable discussions .
10 , phys . lett .
b * 705 * , 174 ( 2011 ) .
. commun . * 64 * , 275 ( 1990 ) .
m. jezabek , et al . ,
. commun . * 70 * , 69 ( 1992 ) .
r. decker , et al .
phys . commun . * 76 * , 361 ( 1993 ) .
et al . , acta phys .
b * 32 * , 1277 ( 2001 ) . n.d .
et al . , arxiv:1002.0543 ( 2010 ) . , comput.phys.commun . * 36 * , 191 ( 1985 ) . , comput.phys.commun . * 79 * , 503 ( 1994 ) . , comput.phys.commun . *
130 * , 260 ( 2000 ) . ,
phys . rev .
d * 41 * , 1425 ( 1990 ) . , j. phys .
g * 37 * , 260 ( 2010 ) .
. commun . * 82 * , 74 ( 1994 ) . ,
j. c part .
fields * 55 * , 553 ( 2008 ) . , phys .
b * 235 * , 198 ( 1990 ) . , atlas - conf-2012 - 009 ( 2012 ) . ,
phys . j. c * 61 * , 33 ( 2009 ) .
, + http://wasm.web.cern.ch/wasm/welcome.html + http://hibiscus.if.uj.edu.pl/@xmath94przedzinski/tau-reweight , arxiv : hep - ph/0508110 ( 2005 ) . ,
* 134 * , 41 ( 2001 ) .
data files generated by a mc event generator ( or constructed using the tau - embedding method ) need to fulfill the following requirements : 1 .
four - momenta of the intermediate boson , the taus and the flavor and the four - momenta of the tau decay products need to be available .
flavor of the intermediate boson needs to be available or set by the user .
3 . for all types of hard processes , but @xmath3 , the four - momenta can be defined in an arbitrary but common frame . for the @xmath3 process , the four - momenta have to be given in the laboratory frame in order to be consistent with the pdfs .
4 . the four - momenta of the taus and their decay products need to be known with sufficient precision in order to ensure numerical stability of the algorithm . six significant digits are recommended .
note that different mc generators may store the truth information in different ways .
it is the responsibility of a user to make sure that all these requirements are fulfilled .
a generic version of the package can be found in @xcite .
the main code is written in c++ and relies upon two libraries : tauola and lhapdf @xcite . a method for reading input information stored using the hepmc @xcite format is prepared .
support for any other input format is available upon request .
* src / tau_reweight_lib.c , src / tau_reweight_lib.h - the core of the algorithm . * src / tauola_wrapper.h - wrapper for tauola fortran routines
. * src / simpleparticle.h - definition of simpleparticle used as a bridge between the event record ( or data file ) and the algorithm .
* src / particle.h - definition of particle used for boosting and rotation of the particles .
* src / read_particles_from_tauola.c , src / read_particles_from_tauola.h - interface to + the hepmc::io_genevent data files used by the example program . *
readme - a short manual .
initialization of tauola .
: : it is ensured by invoking : + tauola::initialize ( ) ; initialization of tauspinner .
: : it is performed by executing : + void initialize_spinner(*bool * ipp , * int * ipol , * double * cmsene ) + where the argument ipp passes the information on the type of collision events ( ipp = true sets @xmath32 collisions ) , ipol passes the information on the spin effects included in the input sample ( ipol= 0 , 1 , 2 corresponds to no spin effects , complete spin effects and spin correlations only , respectively ) and cmsene sets the collision center of mass energy . reading the data files .
: : information on the four - momenta and the flavor of the boson , the final state taus or tau and neutrino pair and the tau decay products is filled and stored in instances of simpleparticle class by the use of the function : + readparticlesfromtauola_hepmc(hepmc::io_genevent & input_file , simpleparticle & boson , + simpleparticle & tau , simpleparticle & tau2 , vector@xmath95simpleparticle@xmath96 & tau_daughters , + vector@xmath95simpleparticle@xmath96 & tau2_daughters ) . + this function should be modified if the input files are not in the hepmc::io_genevent format .
calculation of the spin weight .
: : it is performed by the use of the following functions : + calculateweightfromparticlesworhpn(simpleparticle & boson , simpleparticle & tau , + simpleparticle & tau2 , vector@xmath95simpleparticle@xmath96 & tau_daughters ) for the @xmath4 and + @xmath93 processes + calculateweightfromparticlesh(simpleparticle & boson , simpleparticle & tau , + simpleparticle & tau2 , vector@xmath95simpleparticle@xmath96 & tau_daughters , + vector@xmath95simpleparticle@xmath96 & tau2_daughters ) for the @xmath3 and @xmath93 processes .
attributing tau helicity states .
: : for @xmath3process , the tau helicity states are attributed at the stage of calculation of the spin weight .
the information can obtained by calling gettauspin ( ) function . 1 . the parent boson , the tau , the tau neutrino and the list of tau daughters are identified and boosted to the @xmath1 rest frame in which the tau is aligned along the @xmath45 axis .
the tau daughters are boosted to the tau rest frame .
two angles of spacial orientation of the neutrino from the tau decay , theta2 and phi2 , are calculated and stored .
the tau daughters are rotated by these angles in order to align the neutrino along the @xmath45 axis .
3 . the tauola decay channel is identified .
the tauola fortran subroutine is called to perform calculation of the polarimetric vector @xmath52 .
5 . the polarimetric vector @xmath52 is rotated back using the theta2 and phi2 angles .
the spin weight is calculated using eq .
[ eq:4 ] and returned to the main program . 1 .
the parent boson , the taus and their tau daughters are identified and boosted to the @xmath0 rest frame in which the taus are aligned along the @xmath45 axis .
2 . for each tau : its identified daughters are boosted to its rest frame .
two angles of spacial orientation of the neutrino from the tau decay , theta2 and phi2 , are calculated and stored .
the tau daughters are rotated by these angles to align the neutrino along the @xmath45 axis .
the tauola decay channel is identified .
the tauola fortran subroutine is called to perform calculation of the polarimetric vector @xmath52 .
the polarimetric vector @xmath52 is rotated back using the theta2 and phi2 angles 3 . in case of the @xmath3
decays : the probability @xmath12 is calculated using eqs [ eq:1]-[eq:2 ] .
the spin weight is calculated using eq .
[ eq:6 ] and returned to the main program .
4 . in case of the @xmath93 decays : the spin weight is calculated using eq .
[ eq:7 ] and returned to the main program .
the evolution of the pdf s is invoked from the wrapper for pdf s : + + f(double x , int i d , double ss , double cmsene ) + + where function f calls the evolution function xfx(x , ss , i d ) @xcite .
the pdf sets need to be available locally
. they can be obtained from the lhapdf project website . | final states involving tau leptons are important components of searches for new particles at the large hadron collider ( lhc ) .
a proper treatment of tau spin effects in the monte carlo ( mc ) simulations is important for understanding the detector acceptance as well as for the measurements of tau polarization and tau spin correlations . in this note
we present a tauspinner package designed to simulate the spin effects .
it relies on the availability of the four - momenta of the taus and their decay products in the analyzed data .
the flavor and the four - momentum of the boson decaying to the @xmath0 or @xmath1 pair need to be known . in the @xmath2 case
the initial state quark configuration is attributed from the intermediate boson kinematics , and the parton distribution functions ( pdf s ) .
tauspinner is the first algorithm suitable for emulation of tau spin effects in tau - embedded samples .
it is also the first tool that offers the user the flexibility to simulate a desired spin effect at the analysis level .
an algorithm to attribute tau helicity states to a previously generated sample is also provided .
z. czyczula , t. przedzinski , z.was |
the power conversion efficiency ( pce ) of organic solar cells recently reached 8 % .@xcite the interpenetrating network of donor and acceptor phase is a key issue for further increasing the performance of polymer : fullerene bulk heterojunction solar cells , as it affects the exciton dissociation , charge transport and recombination.@xcite whereas a fine phase intermixing on the sub - nm scale is believed to be beneficial for efficient photogeneration , the charge transport is strongly related to the percolation pathways formed within the bulk of the solar cell .
a crucial parameter having a great impact on the morphology of a bulk heterojunction solar cell is the donor acceptor ratio . for many conjugated polymers , e.g. poly(p - phenylene vinylene ) ( ppv )
@xcite , a fullerene content of 67 wt.-% to 80 wt.-% was found to be optimal for the performance of those solar cells .
only a few polymer : fullerene systems have their optimum blend ratio at 1:1 , e.g. poly(3-hexyl thiophene-2,5-diyl ) ( p3ht):[6,6]-phenyl - c@xmath1-butyric acid methyl ester ( pc@xmath1bm).@xcite recently , it was reported that depending on the chain length and density of the polymer side chains , as well as the size of the fullerene , intercalation of the fullerenes in between the side chains can occur.@xcite mayer et al
. found an optimum polymer : fullerene blend ratio of 1:3 - 1:4 in the blend system consisting of poly(2,5-bis(3-tetradecyl thiophen-2-yl ) thieno[3,2-b]thiophene ) ( pbttt ) : phenyl - c@xmath2-butyric acid methyl ester ( pc@xmath2bm ) , as an excess of pcbm molecules was necessary to create phase segregation . in mdmo - ppv : pcbm blend system the same optimum blend ratio of 1:3 - 1:4 was assigned to a nanoscale phase segregation with pure pcbm domains surrounded by a matrix of polymer containing up to 50 wt.-% of fullerene.@xcite following an efficient charge generation , charge transport and recombination dynamics in blend systems are intuitively expected to be affected by the phase segregation and the dimensions of domain sizes , which will be addressed in detail in this work .
here we report on transport and recombination studies on poly(2,5-bis(3-dodecyl thiophen-2-yl ) thieno[2,3-b]thiophene ) ( pbtct - c@xmath0):pc@xmath1bm blend solar cells in which the ratio between the polymer and the fullerene was varied .
we used two complementary techniques , namely charge carrier extraction by linearly increasing voltage ( photo - celiv ) probing the dynamics on a macroscopic level and time - resolved microwave conductivity ( trmc ) for a microscopic point of view . note that both techniques yield information on the recombination processes very close to open circuit conditions , but on different length scales .
both techniques reveal an enhanced geminate recombination in the 1:1 weight ratio due to a fine donor
acceptor intermixing without phase segregation .
in contrast , in the 1:4 ratio the formation of extended acceptor regions lead to efficient polaron pair dissociation .
the experimental findings are compared with the reference material system p3ht : pcbm .
[ fig : fig1 ] shows x - ray diffraction ( xrd ) measurements on as spun pbtct - c@xmath0 films and those blended with pc@xmath1bm in a 1:1 and 1:4 weight ratio .
no additional annealing was applied . for the pure pbtct - c@xmath0 film a diffraction peak at 4.3@xmath3
is observed , corresponding to a lamellar spacing of 2.0 nm , i.e. the distance between two neighboring sheet - like structures consisting of @xmath4-stacked polymers .
upon addition of pc@xmath1bm , the x - ray peak was shifted towards smaller angles , corresponding to an increase in the lamellar packing distance to 2.8 nm .
no further shift was observed from the 1:1 to the 1:4 weight ratio .
film and blended with pc@xmath1bm in a 1:1 and 1:4 ratio .
the x - ray peak is shifting from 4.45@xmath3 in the pristine pbtct - c@xmath0 film to 3.17@xmath3 in the blend corresponding to a lattice distance of 2.0 nm and 2.8 nm , respectively . ] in order to study the macroscopic charge transport and recombination dynamics in the two blend ratios , we performed photo - celiv measurements at temperatures ranging from @xmath5 k to @xmath6 k. for both pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 devices charge carriers can be extracted from the bulk after laser excitation . however , considering the initial concentration of extracted charge carriers @xmath7 from the photo - celiv measurements determined at a delay time of 150 ns , we observe an order of magnitude difference between the studied blend ratios
. for pbtct - c@xmath0:pc@xmath1bm 1:4 @xmath7 is 9.4@xmath8 m@xmath9 at @xmath5 k , whereas for the 1:1 ratio the initial carrier concentration is decreased to a value of 9.6@xmath10 m@xmath9 . in fig .
[ fig : fig2 ] the time dependent extracted charge carrier concentration @xmath11 at @xmath12 k and @xmath13 k for pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 are compared . for both ratios a bimolecular recombination process can be observed , but with different order of decay . at @xmath13
k the charge carrier decay is faster and more pronounced compared to @xmath12 k for 1:1 and 1:4 blend device , which can be seen in a stronger decrease of the carrier concentration at large delay times .
k and @xmath13 k. the initial extracted charge carrier concentration ( at @xmath14 ns ) is about an order of magnitude less for the 1:1 ratio compared to the 1:4 device . for @xmath15
ns a bimolecular recombination with differing order can be observed for both ratios .
pbtct - c@xmath0:pc@xmath1bm 1:4 reveals a decay order of @xmath16 at @xmath12 k and @xmath17 at @xmath13 k , whereas in the 1:1 ratio an order of @xmath18 and @xmath19 can be determined for @xmath12 k and @xmath13 k. ] using the generalized continuity equation @xmath20assuming that the spatial derivative of the current is zero at the built - in voltage the experimental data can be fitted with a recombination coefficient @xmath21 .
the exponent corresponds to the order of the decay , @xmath22.@xcite from the fits shown in fig .
[ fig : fig2 ] a charge carrier decay order of @xmath18 for 1:1 pbtct - c@xmath0:pc@xmath1bm at @xmath12 k can be obtained , whereas at @xmath13 k the order decreases to @xmath19 . for the 1:4 blend device , the recombination order decreases from 3.3 at @xmath12 k to @xmath17 at @xmath13 k. the average charge carrier mobility in the bulk deduced from photo - celiv measurements at a fixed @xmath23 of 60 @xmath24s is shown in fig .
[ fig : fig4](a ) .
the mobility in the 1:4 blend ratio is about an order of magnitude higher as compared to the 1:1 ratio over the entire studied temperature range with @xmath25 @xmath26v@xmath27s@xmath27 for 1:4 and @xmath28 @xmath26v@xmath27s@xmath27 for 1:1 pbtct - c@xmath0:pc@xmath1bm bhj solar cells at @xmath5 k. note that no charge carriers could be extracted in 1:1 devices below @xmath6 k. furthermore , trmc measurements were performed to investigate the local charge transport in pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 samples .
trmc signals of both samples show a fast initial rise due to the nanosecond laser pulse until the signal decays again . for the 1:4 blend
the maximum photoconductivity @xmath29 at room temperature is more than one order of magnitude higher than that of the 1:1 blend ( see fig .
[ fig : fig3 ] ) , which is in good agreement with the experimentally observed order of magnitude difference in the photo - celiv signal for 1:1 and 1:4 devices . variations of the photoconductivity on a similar scale depending on the relative pc@xmath1bm concentration were observed previously for pc@xmath1bm : poly(phenyl vinylene ) derivatives.@xcite : pc@xmath1bm blends photoexcited at 532 nm and a fluence of 1.4@xmath3010@xmath31 photons @xmath32 per pulse at @xmath5 k. trmc signals were normalized to the optical attenuation of the film .
the inset shows the transients at @xmath12 k fitted by the generalized continuity equation . from the fits a decay order of 2.8 for the 1:4 blend device and 3.1 for the 1:1 ratio was determined .
the laser intensity was adjusted to be 135 @xmath24j@xmath32 per pulse . ]
the decrease of the photoconductivity after @xmath29 is due to charge recombination and due to immobilization caused by trapping of charge carriers .
note that due to the absence of electrodes no charges are collected . in the inset of fig .
[ fig : fig3 ] the product of the quantum efficiency for charge generation @xmath33 and the sum of the mobilities of the positive and negative charge carrier @xmath34 is shown for the 1:1 and 1:4 ratios for @xmath12 k. for the trmc transients we used the same fitting routine as for the photo - celiv transients . in fig .
[ fig : fig4](b ) the product @xmath35 is plotted as a function of @xmath36 for both studied blend ratios .
clearly , two different contributions are observed for the pbtct - c@xmath0:pc@xmath1bm 1:4 sample .
below @xmath12 k a small activation energy of @xmath37 mev can be obtained , whereas for temperatures above 200 k it becomes larger , i.e. @xmath38 mev .
in contrast , only one activation energy of approximately @xmath39 mev can be seen in the 1:1 blend ratio . at @xmath5
k the product @xmath35 for the 1:1 weight ratio is @xmath40 @xmath26v@xmath27s@xmath27 , whereas for the 1:4 blend ratio a value of @xmath41 @xmath26v@xmath27s@xmath27 can be determined .
s and ( b ) @xmath35 deduced from trmc experiment is plotted against @xmath36 for pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 . from trmc
two activation energies can be observed in the 1:4 blend ratio52 mev for @xmath42 k and 19 mev for @xmath43 k , whereas in the 1:1 ratio only one activation energy of 13 mev over the whole temperature range is detected ] the recombination order deduced from the fit of the photo - celiv and trmc transients for pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 bhj solar cells is shown in fig .
[ fig : fig5 ] as a function of temperature . from both techniques
an increasing recombination order with decreasing temperature can be determined for both studied blend ratios . in the 1:4 device a charge carrier decay order of @xmath44 was deduced from trmc and @xmath19 from photo - celiv at @xmath5 k increasing to around 3.4 and 4.6 at @xmath6 k , respectively .
similar decay orders were observed in 1:1 blend devices with an increase from 2.1 to 4.9 obtained from the photo - celiv experiments and 2.3 to 3.1 from corresponding trmc transients .
we note that slightly higher recombination orders are revealed from photo - celiv measurements at low temperatures , the origin of which will be discussed . for comparison
, the data of an annealed p3ht : pc@xmath1bm 1:0.8 device is also shown . there ,
the experimentally determined recombination order increases only slightly from 2.1 to 2.2 in the studied temperature range from @xmath5 k to @xmath6 k , respectively .
data of the photo - celiv and @xmath45 data of the trmc transients for pbtct - c@xmath0:pc@xmath1bm 1:1 and 1:4 . for comparison the blend system p3ht : pc@xmath1bm is also shown . ]
in x - ray diffraction measurements a lamellar spacing of 2.0 nm can be found in a pure pbtct - c@xmath0 film .
this is slightly larger than the 1.84 nm spacing peak observed for a thermally annealed film,@xcite but is nonetheless shorter than the length of two dodecyl chains placed end - to - end , which is indicative of some interdigitation of the side chains of adjacent polymer backbones.@xcite for both pbtct - c@xmath0:pc@xmath1bm blend ratios a clear shift of the xrd peak is observed pointing at an increase of the lamellar distance of 0.8 nm due to incorporation of methanofullerenes in the studied blend system .
similar to ppv : pc@xmath1bm blend systems@xcite it was suggested for pbttt : pc@xmath2bm devices chemically related to the blend system studied in this work that the addition of more than 50 wt.% of fullerene lead to phase segregation with pc@xmath2bm domains growing in proportion to an increasing amount of fullerene.@xcite thus , although residual pc@xmath1bm molecules in the 1:4 blends are not contributing to a further increase of the lamellar distance , the changed phase segregation does affect the transport properties and recombination dynamics in a bulk heterojunction solar cell , which was already suggested previously@xcite and will be discussed in the following . in the 1:4 blend device
a higher charge carrier mobility in the bulk ( photo - celiv ) as well as detected locally ( trmc ) is observed experimentally . from temperature
dependent trmc measurements the contribution of charge carriers to the photoconductance can be deduced by determining the activation energy from an arrhenius plot . in pbtct - c@xmath0:pc@xmath1bm 1:4 two different activation energies could be assigned ( see fig . [
fig : fig4 ] ) with a transition temperature of @xmath12 k. similar results were reported previously in different polymer : pc@xmath1bm blend systems , e.g. p3ht : pc@xmath1bm.@xcite there , the activation energy for @xmath42 k was assigned to the contribution of electrons in pc@xmath1bm , whereas at low temperatures the hole mobility in the polymer dominates the product @xmath35 .
if we project these findings to the here studied pbtct - c@xmath0:pc@xmath1bm solar cells , above @xmath12 k we observe only an electron mobility in the 1:4 blends .
in contrast , in the 1:1 blend ratio holes on pbtct - c@xmath0 dominate @xmath35 over the entire studied temperature range .
note that from photo - celiv measurements the polarity of the extracted charge carriers can not be distinguished .
however , we can conclude that in the 1:4 blend ratio extensive phase segregation leads to an efficient charge transport , whereas both charge carriers contribute to the transport .
in contrast , a reduced charge carrier mobility is observed in the 1:1 pbtct - c@xmath0:pc@xmath1bm bhj solar cell , which we assign to the fine donor
acceptor intermixing with less percolation paths for the polarons to the corresponding electrodes .
furthermore , from current voltage characteristics of the two blend ratios we find an almost three times higher short circuit current density under am1.5 g illumination in the 1:4 blend ratio compared to the 1:1 ratio .
the initial charge carrier concentration experimentally found within ( trmc experiment , fig .
[ fig : fig3 ] ) and extracted from ( photo - celiv , fig .
[ fig : fig2 ] ) the device for both pbtct - c@xmath0:pc@xmath1bm blend ratios was about an order of magnitude lower at 150 ns in the 1:1 device as compared to 1:4 with both techniques .
this was despite the higher polymer content and thus larger absorption coefficient at 532 nm in the former .
we note that the trmc signal was corrected for the optical attenuation ( see equation [ eq : g2 ] ) .
photoluminescence in intercalated bulk systems is quenched very efficiently,@xcite indicating that almost every photogenerated singlet exciton is separated at a donor
acceptor interface creating a polaron pair .
however , fine phase intermixing and disorganized percolation paths in pbtct - c@xmath0:pc@xmath1bm 1:1 blends may lead to a less efficient separation of photogenerated polaron pairs as compared to the 1:4 ratio .
thus , we assign the reduced charge carrier extraction in the 1:1 device at @xmath46 ns to an enhanced geminate recombination occurring on shorter time scales due to lower delocalization of the charge carriers caused by a lack of percolation paths .
electrons may be trapped on isolated acceptor molecules , which can not hop away from the interface to be dissociated .
consequently , those charge carriers are lost for the photocurrent . during the delay time
the recombination of charge carriers will change the flat band conditions of the solar cell resulting in a sweep out of some of the charges , in addition to recombination .
this effect is proportional to the charge carrier mobility and therefore stronger at higher temperatures .
therefore , a lower @xmath7 is observed in photo - celiv measurements at high temperatures for a given donor
acceptor ratio ( see fig . [
fig : fig2 ] ) . in the following ,
we will discuss the charge carrier decays after 150 ns in view of polaron recombination .
the crucial role of charge trapping for recombination orders above two will be discussed .
recombination of charges in low mobility materials is usually described by the bimolecular langevin process.@xcite this mechanism shows a charge carrier decay of the second order if electron and hole concentrations are similar . in typical donor
acceptor blends , the transport of electrons is restricted to the acceptor and the hole transport to the donor domain . in these systems ,
recombination of electrons and holes is spatially restricted to the heterointerface,@xcite which explains in part the overestimation of recombination rates by applying the langevin model.@xcite in addition to reducing the overall recombination probability , it is important to note that the occurrence of non geminate recombination described by langevin theory requires either mobile electrons and holes , @xmath47 and @xmath48 respectively , to meet at the heterointerface , or mobile electrons ( holes ) and holes ( electrons ) trapped very close to the heterointerface .
thereby , we distinguish between spatially @xmath49 and energetically @xmath50 trapped carriers .
the former is related to charge carriers being spatially confined in a material phase due to the lack of percolation to the respective electrode .
in contrast to energetically trapped charges , they are locally mobile and can thus recombine at the heterointerface with an oppositely charged carrier , but can not be extracted at the contacts . energetically trapped charges which are immobile can only participate in the recombination process if ( i ) they are close enough to the donor
acceptor interface ( for instance due to donor
acceptor intermixing ) , or ( ii ) if they are emitted after a certain dwell time , thus being mobile once again ( @xmath51 ) . for ( ii )
, the emission rate @xmath52 of charges from traps depends crucially on the energetic distribution of the density of trap states . in the hopping systems under consideration ,
even the intrinsic density of states below the transport energy level the tail of an exponential or a gaussian density of states distribution ( dos)can act as traps.@xcite if we consider the continuity equation of free holes @xmath47 , which can be extracted , we find @xmath53 . here ,
@xmath48 and @xmath54 are the free and trapped electrons , respectively , and @xmath21 and @xmath55 are the relevant recombination prefactors . if @xmath56 , the recombination dynamics become @xmath57 , where @xmath58a first order process .
depending on the magnitude of @xmath54 as compared to the mobile charge carriers , recombination orders between one and two would be observed . in the studied pbtct - c@xmath0:pc@xmath1bm blend system a recombination order far above two is observed at low temperatures by photo - celiv and trmc for both blend ratios ( fig .
[ fig : fig5 ] ) approaching close to or even decrease below two at @xmath5 k. in 1:4 blends , the excess amount of fullerenes leads to the formation of acceptor rich domains , which allows a faster delocalisation on short time scales improving the photogeneration yield @xcite and to percolation paths for the electrons , favoring the charge extraction process as compared to the 1:1 blend .
due to fine intermixing in the 1:1 ratio , electrons will be mostly trapped , if not energetically , then spatially due to the lack of percolation paths on isolated fullerenes close to the polymer backbones . a large amount of polaron pairs recombines before dissociation , as indicated by the low initial signal magnitude for both , trmc ( fig .
[ fig : fig3 ] ) and photo - celiv ( fig .
[ fig : fig2 ] ) measurements .
the remaining charge carriers might find percolation paths to be extracted .
however , with less thermal energy available they will partly be trapped in deep tail states of the dos , thus being immobile . as emission is thermally activated
, the release of trapped charges takes much longer at low temperatures . only after emission from the continuous distribution of trap states , the now mobile charges
can either be extracted by the linear voltage pulse or recombine non geminately .
the latter process becomes less probable the later the emission occurs : the recombination partners of opposite charge may already have recombined or left the device .
this leads to a reduced charge carrier decay .
therefore , we assign the observed high recombination order at low temperatures to a delayed release of trapped charges not actively participating in the recombination process.@xcite with increasing temperature energetic trapping is more unlikely . however , the spatial restrictions in the bulk structure will affect the charge dynamics more dominantly .
the recombination order decreases even slightly below two in both studied ratios , which we attribute to an imbalanced relation between free mobile electrons and holes due to spatial trapping in the fine intermixed phase , as explained above .
even though the two complementary techniques photo - celiv and trmc rely on different measurement principles , the results fit qualitatively together . in case of trmc ,
all mobile charge carriers contribute to the measurement signal , including also spatially restricted charge carriers @xmath49 .
photo - celiv is only sensitive to charge carriers , which can be extracted by the voltage pulse .
this includes also charge carriers being energetically trapped in deep states during the delay time not participating in the recombination processes , but can be extracted by the voltage pulse .
thus a higher recombination order is deduced from the photo - celiv experiment .
the transport and recombination dynamics in pbtct - c@xmath0:pc@xmath1bm bulk heterojunction solar cells were investigated by the combination of the two complementary techniques of photo - celiv and trmc for the first time probing different length scales .
x - ray diffraction measurements indicate that structural changes occurred in the studied blend structure for a 1:1 and 1:4 weight ratio , where pc@xmath1bm molecules probably intercalate into the available space of two neighboring polymer side chain stacks . due to fine donor acceptor intermixing ,
an enhanced geminate recombination is explained by the order in magnitude less concentration of initially extracted polarons in the 1:1 ratio . the lack of percolation pathways for electrons in the 1:1 ratio results in a low local charge carrier mobility decreasing the polaron pair dissociation probability significantly .
in contrast , in the 1:4 ratio extensive phase segregation lead to an efficient charge generation as the polarons can easily hop away from the heterointerface .
the polaron recombination dynamics studied by trmc and photo - celiv are bimolecular with an increasing order with decreasing temperature for both studied ratios , which is due to charge trapping .
bhj solar cells were prepared by spin coating a 35 nm layer of poly(3,4-ethylene dioxy thiophene):poly(styrene sulfonate ) ( baytron p vp ai 4083 ) on indium tin oxide samples with post - annealing step of 130@xmath3c for 10 minutes .
the pbtct - c@xmath0:pc@xmath1bm blends made from solutions of 16 mg / ml and 10 mg / ml in 1,2-ortho - dichlorobenzene were spin coated in an inert atmosphere with active layers of a thickness in the range of 70 to 280 nm . without further annealing the active blend layer , ca ( 3 nm)/al ( 100 nm ) were evaporated thermally on top .
pc@xmath1bm was purchased from solenne , poly(2,5-bis(3-dodecylthiophen-2-yl)thieno[2,3-b]thiophene ) ( pbtct - c@xmath0 ) was synthesized according to the published procedure in ref .
the molecule weight was mn 22,600 g / mol mw 39,300 g / mol as determined by gpc in chlorobenzene at 60@xmath3c on an agilent 1100 series hplc using two polymer laboratories mixed b columns in series , with calibration against narrow weight pl polystyrene calibration standards .
all materials were used without further purification .
the current voltage characteristics of the organic solar cells were measured in a nitrogen glovebox .
we used an oriel 1160 am1.5 g solar simulator for illumination . for investigations of the charge transport and recombination studies using the photo - celiv technique ,
the solar cells were transferred to a closed cycle helium cryostat with helium as inert contact gas .
a triangular voltage pulse in reverse direction is applied to the solar cells extracting free charge carriers from the bulk . when exciting the sample with a laser flash , photogenerated charge carriers can be studied .
@xcite by applying an offset voltage to compensate for the built - in field of the solar cells and varying the delay time between the laser excitation and the charge carrier extraction , the recombination dynamics can be studied by measuring the time dependent charge carrier concentration . using a parametric equation for the mobility @xcite @xmath59 the average charge carrier mobility in the bulk
can be calculated from the transient , with @xmath60 , @xmath61 the sample thickness , @xmath62 the time of the transient maximum , @xmath63j as the maximum of the extraction current and @xmath64 the current step related to the capacitive displacement current.@xcite a double pulse generator ( agilent 81150a ) was used for applying a triangular voltage pulse to the solar cell .
the current transients were acquired by a digital oscilloscope ( agilent infiniium dso90254a ) after amplification by a current voltage amplifier ( femto dhpca-100 ) .
photo - celiv measurements were performed at different temperatures ranging from 150 k to 300 k in steps of 25 k. the delay time between the laser excitation and the extraction voltage pulse was varied from 100 ns to 10 ms .
we used the second harmonic of a nd : yag laser ( @xmath65 nm , @xmath66 80 ps pulse duration ) for laser excitation .
furthermore , the decay of charge carriers was studied by transient microwave conductivity . by using this technique both mobile holes and electrons
contribute to the measured change in microwave conductance ( @xmath63 g ) .
the microscopic charge carrier mobilities were deduced from the trmc transients by equation @xmath67 @xmath33 is the quantum efficiency for charge generation and @xmath34 is the sum of the mobilities of the positive and negative charge carrier .
@xmath68 denotes the ratio between the broad and narrow inner dimension of the waveguide , @xmath69 is the elementary charge , @xmath70 the incident laser fluence and @xmath71 is the fractions of photons absorbed by the sample .
the laser fluence in the trmc experiment was 135 @xmath24j@xmath32 per pulse . as described in more detail
previously @xcite , if @xmath33 is undetermined , the value of @xmath45 is a lower limit of the local charge carrier mobility .
samples for trmc measurements were prepared from the same solution as the solar cells .
the current work is supported by the bundesministerium fr bildung und forschung in the framework of the grekos project ( contract no .
03sf0356b ) .
thanks the german federal environmental foundation ( deutsche bundesstiftung umwelt , dbu ) for funding .
the work of d.h.k.m . forms part of the research programme of the dutch polymer institute ( dpi ) , project nr .
c.d . gratefully acknowledges the support of the bavarian academy of sciences and humanities .
v.d.s work at the zae bayern is financed by the bavarian ministry of economic affairs , infrastructure , transport and technology .
i. mcculloch , m. heeney , m. l. chabinyc , d. delongchamp , r. j. kline , m. clle , w. duffy , d. fischer , d. gundlach , b. hamadani , r. hamilton , l. richter , a. salleo , m. shkunov , d. sparrowe , s. tierney , and w. zhang , , * 2009 * , , 1091 . | we studied the recombination dynamics of charge carriers in organic bulk heterojunction solar cells made of the blend system poly(2,5-bis(3-dodecyl thiophen-2-yl ) thieno[2,3-b]thiophene ) ( pbtct - c@xmath0):[6,6]-phenyl - c@xmath1-butyric acid methyl ester ( pc@xmath1bm ) with a donor acceptor ratio of 1:1 and 1:4 . the techniques of charge carrier extraction by linearly increasing voltage ( photo - celiv ) and , as local probe , time - resolved microwave conductivity ( trmc ) were used .
we observed a difference in the initially extracted charge carrier concentration in the photo - celiv experiment by one order of magnitude , which we assigned to an enhanced geminate recombination due to a fine interpenetrating network with isolated phase regions in the 1:1 pbtct - c@xmath0:pc@xmath1bm bulk heterojunction solar cells .
in contrast , extensive phase segregation in 1:4 blend devices leads to an efficient polaron generation resulting in an increased short circuit current density of the solar cell . for both
studied ratios a bimolecular recombination of polarons was found using the complementary experiments .
the charge carrier decay order of above two for temperatures below 300 k can be explained by a release of trapped charges .
this mechanism leads to a delayed bimolecular recombination processes .
the experimental findings can be generalized to all polymer : fullerene blend systems allowing for phase segregation .
this is the pre - peer reviewed version of the following article : influence of phase segregation on the recombination dynamics in organic bulk heterojunction solar cells , a. baumann , t. j. savenije , d. h. k. murthy , m. heeney , v. dyakonov , c. deibel , which has been published in final form at + * http://onlinelibrary.wiley.com/doi/10.1002/adfm.201002358/abstract[adv .
func . mat .
21 , 1687 ( 2011 ) ] * |
the fastest variability components in x - ray binaries , the kilohertz quasi - periodic oscillations ( khz qpos ) , have been detected in about thirty neutron star low - mass x - ray binaries ( ns lmxbs ) since first discovered with rossi x - ray timing explorer in 1996 ( see van der klis 2006 for a review ) .
twin khz qpos appeared simultaneously in about twenty ns lmxbs .
these qpos are thought to reflect the motion of matter at the inner edge of an accretion disc around the ns .
early observations showed that the frequency separation ( @xmath0 ) of the twin khz qpos is close to the ns spin frequency ( @xmath1 ) ( e.g. strohmayer et al .
1996 ; ford et al .
1997 ) , suggesting a beat - frequency explanation ( miller , lamb & psaltis 1998 ) .
however , the more detailed measurements revealed that @xmath0 is generally inconsistent with a constant value of @xmath1 but varying with the upper ( @xmath2 ) or lower ( @xmath3 ) frequency of the twin khz qpos ( van der klis et al .
1997 ; mndez et al .
1998 , 1999 ) .
stella & vietri ( 1999 ) propose a relativistic precession model which predicts a changing peak separation of the twin khz qpos .
the upper and the lower khz qpo frequencies are identified with the keplerian frequency of the rotational plasma flow at the inner edge of the disc and the periastron precession frequency of the orbit , respectively . in this model ,
the ns spin frequency plays no role in setting up the frequencies of the khz qpos , but a massive ( @xmath4 ) ns is usually required to match the observations .
abramowicz et al . ( 2001 , 2003 ) suggest that the twin khz qpos can be explained by a nonlinear resonance in the epicyclic motion in the accretion disc , leading to the @xmath5 ratio of the upper and lower frequencies , although whether there is an intrinsically preferred ratio between the @xmath2 and @xmath3 is controversial ( belloni et al .
2007 ) .
the subtle feature of twin khz qpos is that @xmath6 is variable but seems to be around either @xmath1 or @xmath7 ( smith , morgan & bradt 1997 ; wijnands & van der klis 1997 ; markwardt , strohmayer & swank 1999 ; chakrabarty et al 2003 ; wijnands et al .
2003 ; barret , boutelier & miller 2008 ; see , however , yin et al . 2007 ; mndez & belloni 2007 ) .
this has motivated some scenarios of khz qpos taking into account the effect of the ns spin .
osherovich & titarchuk ( 1999 ) suggest that khz qpos could be modeled as oscillations of the blobs thrown into magnetosphere from the inner edge of the accretion disc .
the lower khz qpo frequency @xmath3 is identified as the keplerian frequency in the disc and the upper frequency @xmath2 the hybrid frequency about the keplerian frequency and the ns spin frequency .
lee , abramowicz & kluiak ( 2004 ) show that the khz qpos may be attributed to forcing of epicyclic motions in the accretion disc by the ns , which induces resonance at selected frequencies when the frequency separation @xmath6 is equal to @xmath1 or @xmath7 .
li & zhang ( 2005 , see also zhang 2004 ) present an alternative interpretation for the origin of the twin khz qpos by considering the interaction between the ns magnetic field and the surrounding accretion disc , which gives rise to mhd loop oscillations at the inner edge of the disc with the frequencies depending on the spin frequencies . in this paper
, we propose a model to explain khz qpos in ns lmxbs based on the interaction of accreting plasma with the ns magnetosphere .
this model is partially based on the interpretation of zhang ( 2004 ) and rezania & samson ( 2005 ) . in the latter work
it was argued that distortion of the ns magnetosphere by the infalling plasma of the keplerian accretion flow can excite resonant shear alfvn waves in a region of enhanced density gradients , where accretion material flows along the magnetic field lines in the magnetosphere .
the major difference between our model and rezania & samson s ( 2005 ) is that we have included the effect of the gravity and spin of the ns on the field line resonances .
this paper is organized as follows . in section 2
we introduce the basic physical model and derive the relation between @xmath1 , @xmath3 and @xmath2 . in section 3
we compare the theoretical relations with the observational data of six sources ( 4u 0614@xmath809 , 4u 1608@xmath952 , 4u 1636@xmath953 , 4u 1728@xmath934 , 4u 1915@xmath905 , and xte 1807@xmath9294 ) . in section 4
we summarized our results and discuss their possible implications .
in lmxbs the plasma from the donor star accretes onto the ns via an accretion disc .
material in the disc firstly rotates in a keplerian motion , then corotates with the magnetosphere after it is trapped completely by the ns magnetic field at the magnetospheric radius , and finally flows along the field lines to the polar cap .
some resonant modes may be excited by the perturbations at the magnetospheric radius when the plasma begins to corotate with the magnetosphere ( osherovich & titarchuk 1999 ; lee et al .
2004 ; zhang 2004 ; rezania & samson 2005 ) .
we consider the qpos as a modulation effect of the mhd waves which are produced at the magnetospheric radius , and the coupling of the two resonant mhd modes may lead to the twin khz qpos in the power spectrum .
we consider the mhd equations in a frame of reference corotating with the ns ( shown in fig . 1 ) , written as follows ( landau & lifshitz 1976 ) @xmath10 @xmath11 @xmath12 @xmath13 where @xmath14 is the plasma velocity , @xmath15 electric current , @xmath16 magnetic field , @xmath17 the displacement from the center of the ns to the plasma , @xmath18 plasma density , @xmath19 barometric pressure , @xmath20 adiabatic index , @xmath21 gravitational constant , @xmath22 and @xmath23 the mass and the angular velocity of the ns is the angular velocity of the ns magnetosphere , which could be slightly deviate from that of the ns .
] , respectively .
the third , fourth and fifth terms on the rhs of eq .
( 1 ) represent the coriolis force , the centrifugal force and the gravity , respectively .
observationally the accretion rates in lmxbs change on a timescale of @xmath24 .
this is much more than the relaxation ( or dynamic ) timescale of the plasma at the inner disc radius ( @xmath25 ) .
so we can approximate the plasma to be in an equilibrium state , which is subject to small perturbations , as discussed in benz ( 2002).by use of the current expression , @xmath26 where @xmath27 is is vacuum magnetic conductivity , eqs .
( 1)-(3 ) can be transformed to be @xmath28 @xmath29 @xmath30 where the subscript 0 denotes variables in the equilibrium state .
the initial relative velocity ( @xmath31 ) is equal to zero in the corotating reference system and can be expressed as @xmath32 in the inertial reference system .
now we consider the mhd equations for the plasma subject to small perturbations , @xmath33 @xmath34 @xmath35 @xmath36 where @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 with the subscript s denoting the perturbed quantities ( @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 ) and with the superscript @xmath47 the variables after the disturbance .
combining eqs .
( 6)-(12 ) we get the equations about the perturbed quantities in the first order approximation , @xmath48 + 2\rho _ 0 { { \bm v } } _ { { \bm s } } \times { { \bm \omega } } + \rho _ 0 \omega ^2 { { \bm r}}_{{\bm s}}\\ & & - \rho _ 0 ( { { \bm \omega } } \cdot { { \bm r } } _ { { \bm s } } ) { { \bm \omega } } - \rho _ 0 \textstyle{{gm } \over { r_0 ^3 } } { { \bm r } } _ { { \bm s } } , \end{array}\ ] ] @xmath49 and @xmath50 differentiating eq . ( 13 ) and substituting @xmath51 into it give @xmath52\\ & & + 2\rho _ 0 \textstyle{\partial \over { \partial t } } ( { { \bm v } } _ { { \bm s } } \times { { \bm \omega } } ) + \rho _ 0 \omega ^2{{\bm v } } _ { { \bm s } } - \rho _ 0 ( { { \bm \omega } } \cdot { { \bm v}}_{{\bm s } } ) { { \bm \omega } } - \rho _ 0 \textstyle{{gm } \over { r_0 ^3}}{{\bm v } } _ { { \bm s}}. \end{array}\ ] ] we assume that ( 1 ) the accretion disc is infinitesimally thin , ( 2 ) the magnetic moment and the spin of the ns are parallel to the @xmath53 axis , and normal to the disc , i.e. , @xmath54 and @xmath55 close to the inner edge of the disc , and ( 3 ) the @xmath56 and @xmath57 axes are along the disc plane , and the mhd wave is assumed to propagate in the @xmath58 plane , i.e. , the wave vector @xmath59 , where @xmath60 is the angle between the @xmath53 axis and @xmath61 . after carrying out fourier transformation for eqs .
( 14)-(16 ) we get the following dispersion equations , @xmath62 @xmath63 @xmath64 where @xmath65 is imaginary unit , @xmath66 ( @xmath67 ) , @xmath68 ( @xmath69 ) , @xmath70 ( @xmath71 ) , and @xmath72 are alfvn velocity , acoustic velocity , keplerian angular velocity , and angular velocity of the perturbation at @xmath73 respectively , and @xmath74 , @xmath75 , @xmath76 are the three components of the perturbed quantity of the speed . equations ( 17)-(19 ) show that there exist three resonance mhd modes . at the magnetospheric radius @xmath73 , the magnetic energy density is equal to the total kinetic energy density , i.e. @xmath77 ( davidson & ostriker 1973 ; ghosh et al .
since the characteristic wavelength is in the same order with the magnetospheric radius ( rezania & samson 2005 ) , we then have @xmath78 , or @xmath79 . because the thermal pressure of the plasma might be comparable with the magnetic pressure ( @xmath80 ) just inside the magnetosphere ( miller et al .
1998 ) , we also suppose @xmath81 . here both @xmath82 and @xmath83 are taken to be constant for certain sources .
substitute these relations into eqs .
( 17)-(19 ) we can get the resonant modes . specifically when @xmath84 , from eqs .
( 17 ) and ( 18 ) we can get @xmath85 , i.e. @xmath86 . substituting this relation into the eqs .
( 17)-(19 ) can give @xmath87 @xmath88 @xmath89 where the negative solutions are excluded .
similarly when @xmath90 , @xmath91 @xmath92 @xmath93 note that in the latter case the mhd wave could nt propagate very far in the accretion disc , so we consider the coupling modes that propagate along the @xmath53 axis ( @xmath94 ) as a more promising explanation of the qpos .
we first rule out the possibility of the @xmath95 mode as the lower khz qpos , since eq .
( 21 ) requires that it should always higher than @xmath96 for stable accretion , which is contradicted with observations .
the coupling between @xmath97 and @xmath98 can also be excluded , which implies a constant frequency separation .
so we suggest the upper and the lower khz qpo frequencies be @xmath99 and @xmath100 . from eqs .
( 20 ) and ( 21 ) we can get the following relation between the frequencies of the upper and the lower khz qpos , @xmath101 where @xmath102 , or @xmath103 @xmath104 where @xmath105 and @xmath106 .
equations ( 27 ) and ( 28 ) indicate that the twin khz qpos may be divided into two groups , with the slope of the @xmath107 vs. @xmath108 relation either larger or smaller than 1 . in the following we call them the large slope coefficient sources ( lscs ) and the small slope coefficient sources ( sscs ) , respectively .
we compare in fig . 2 the @xmath107 vs. @xmath108 relations obtained in last section with the observed khz qpos in six sources 4u 0614@xmath809 , 4u 1608@xmath952 , 4u 1636@xmath953 , 4u 1728@xmath934 , 4u 1915@xmath905 , xte 1807@xmath9294 , in which both the spin and twin khz qpo frequencies have been measured .
the spin frequencies , disposed in table 1 , are from van der klis ( 2006 ) , mndez & belloni ( 2007 ) , yin et al .
( 2007 ) , altamirano et al .
( 2008 ) , and their references . for 4u 0614 + 09
we adopt the updated spin frequency of 415 hz ( strohmayer , markwardt & kuulkers 2008 ) .
the dots with error bars represent the measured values , and the solid lines stand for theoretical relations .
we distinguish the @xmath107 vs. @xmath108 relations for sscs and lscs , and accordingly adopt relation ( 27 ) to fit the data for 4u 0614@xmath809 , 4u 1608@xmath952 , 4u 1636@xmath953 , and 4u 1728@xmath934 , and relation ( 28 ) for the other two sources , 4u 1915@xmath905 and xte 1807@xmath9294 .
for each source , the value of @xmath109 or @xmath110 for best fitting is also shown in the figure .
it is noted that a cluster of the values ( @xmath111 ) of @xmath109 and @xmath110 can well reproduce the observed relations . in the left panel of fig .
3 we show the observed and predicted relations for all of the six sources . in the right panel
we plot the relation between @xmath112 and @xmath108 by use of the parameter @xmath113 or @xmath110 that we have got . for sscs the peak separation of the twin khz qpos is less than the spin frequency , i.e. , @xmath114 , and decreases with @xmath3 or @xmath2 ; for lscs the peak separation is more than the spin frequency , i.e. , @xmath115 , and increases with the increasing @xmath3 or @xmath2 . in the former group , @xmath116 is around @xmath1 for 4u 0614@xmath809 and 4u 1728@xmath934 , and @xmath7 for 4u 1608@xmath952 and 4u 1636@xmath953 ( miller et al .
1998 ; van der klis 1997 ; stella , vietri & morsink 1999 ; lewin & van der klis 2006 ; mendez & belloni 2007 ) .
our final note is that when the alfven speed is equal to the acoustic speed of the plasma , i.e. @xmath117 , eq .
( 26 ) will recover to the expression in the sonic - point beat - frequency model ( miller et al .
1998 ) . in this case
the peak separation is equal to the spin frequency and almost invariant .
in this paper we propose a resonant mhd model for the twin kilohertz qpos in lmxbs .
the modes of the mhd waves vertical and parallel to the accretion disc are derived , and the twin khz qpos frequencies are identified with the frequencies of the two resonant modes . in this model
the twin khz qpo frequencies are correlated with the spin frequencies , and the separation frequencies also change with the qpo frequencies .
we show that the measured relations between @xmath3 , @xmath2 , and @xmath1 can be accounted for with reasonable values of the input parameters .
there are several spin - involved mhd models for khz qpos in the literature .
osherovich & titarchuk ( 1999 ) suggest that khz qpos can be explained as oscillations of large scale inhomogeneities ( hot blobs ) thrown into the ns magnetosphere .
participating in the radial oscillations with keplerian frequency , such blobs are simultaneously under the influence of the coriolis force .
the derived frequency relation is @xmath118 , or @xmath119 , which is plotted in the dotted curve in fig .
the significant deviation from the measured data indicates that this model is not successful for most of the lmxbs . in their mhd loop oscillation model
li & zhang ( 2005 ) derive a linear frequency relation , i.e. , @xmath120 where @xmath121 is an input parameter . here
the upper khz qpo frequency is assumed to the keplerian frequency and the lower khz qpo is identified as the principal fast kink mode of the standing mhd waves along the toroidal field lines at the magnetospheric radius . the relation is also plotted in fig . 4 in solid curves .
a comparison with the measured data shows that the fit is acceptable for all the six sources .
rezania & samson ( 2005 ) propose a model for qpos in lmxbs based on oscillating magnetohydrodynamic modes in ns magnetospheres .
they argue that the interaction of the accretion disc with the magnetosphere can excite resonant shear alfvn waves in a region of enhanced density gradients , where accretion material flows along the magnetic field lines in the magnetosphere .
the predicted @xmath122 ratio is found to be independent of the ns spin frequencies .
the main difference between rezania & samson ( 2005 ) and this work lies in that rezania & samson ( 2005 ) assume that the strong gravity of the ns produces a converging flow which will hit the star s magnetosphere in a large velocity , while we consider the motion of the plasma is still mainly keplerian , before they enter the magnetosphere and corotate with the magnetosphere .
we also include the effect of the gravity and the rotation of the ns for the trapped plasma , and find that the resulting wave frequencies relate with both the spin frequency and the gravity .
so far we have assumed that the spin and magnetic axes of the ns are aligned in lmxbs .
however , the existence of persistent millisecond pulsations in some lmxbs indicates that the magnetic axis is at least somewhat tilted from the spin axis . in this case
the magnetic field is no longer homogeneous at a given radius in the accretion disc , and the orbit of plasma in the inner region of the disc becomes non circular .
furthermore , the stellar magnetic field can induce disk warping and precession ( e.g. lai 1999 , 2003 ) , and modulate the orbit and hence the qpo frequencies . for x - ray binaries
the disc precession timescale is usually of tens of days to years ( cf .
wijers & pringle 1999 and references therein ) , which is much longer than the duration of each qpo observation . additionally for accretion - powered millisecond pulsars
the magnetic inclination is likely to be very small ( lamb et al . , 2008 ) .
for the above reasons we expect that the change of the qpo frequencies induced by oblique magnetic fields might be very small compared with the uncertainties in the measured frequencies .
our results indicate that the peak separation is always related to the spin frequency .
what s more , there seems to be a weak positive correlation between the spin frequency and the parameter @xmath109 for sscs , which can be described as @xmath123 , plotted in the left panel of fig .
5 . substitute this relation into @xmath124 we get a trend of @xmath125 changing with the spin frequency , as plotted in the right panel of fig .
5 . from the dark black curve to the light gray curve @xmath3 changes from 1000 hz to 100 hz in a step of 100 hz .
we find that when @xmath1 increases , @xmath116 varies from @xmath126 to @xmath127 , and finally to @xmath128 .
the transitions occur at @xmath129 hz and 500 hz , respectively .
the accretion process can take place only when the magnetospheric radius is less than the corotation radius ( e.g. ghosh & lamb 1979 ) .
in other words , the keplerian frequency at the magnetosphere radius should be more than the spin frequency of the ns if accretion process can take place . because of this there is a minimum value of the lower frequency of the twin khz qpos in sscs , @xmath130 .
besides , due to the fact that the peak separation must be positive values we can get the maximal value for the upper frequency , @xmath131 for sscs . these may serve as possible evidence to testify this model with future measurements of khz qpos in lmxbs .
the authors thank the anonymous referee for the helpful suggestion on the manuscript .
this work was supported by the natural science foundation of china under grant numbers 10573010 and 10221001 .
99 abramowicz m.a . , kluziak w. , 2001 , a&a , 374 , l19 abramowicz m.a . , bulik t. , bursa m. , kluziak w. , 2003 , a&a , 404 , l21 altamirano d. , van der klis m. , mndez m. , jonker p.g . ,
klein - wolt m. , lewin w. h. g. , 2008 , apj , accepted ( astro - ph/0806.0962 ) barret d. , boutelier m. , miller m.c . , 2008 , mnras , 384 , 1519b belloni t. , homan j. , motta s. , ratti e. , mndez m. , 2007 , mnras , 379 , 247b benz a.o . , 2002 , plasma astrophysics : kinetic processes in solar and stellar coronae , 2nd edition , astrophysics and space science library , vol .
279 , kluwer academic publishers , dordrecht boirin l. , barret d. , olive j.f . , bloser p.f . , grindlay j.e . , 2000 ,
a&a , 361 , 121 boutloukos s. , lamb f.k.,2008 , aipc , 983 , 533 boutloukos s. , van der klis m. , altamirano d. , klein - wolt m. , wijnands r. , jonker p.g . , fender r.p . , 2006 , apj , 653 , 1435b chakrabarty d. , morgan e.h . ,
muno , m.p .
, galloway d.k . , wijnands r. , van der klis m. , markwardt c.b . 2003 , nature , 424 , 42 davidson k. , ostriker j.p .
1973 , apj , 179 , 585 di salvo , t. , mndez m. , van der klis m. , ford e. , robba n.r . , 2001 , apj , 546 , 1107 di salvo t. , mndez m. , van der klis m. , 2003 , a&a , 406 , 177 ford e.c .
et al . , 1997 , apj , 486 , 47 galloway d.k . , chakrabarty d. , muno m.p . , michael p. , savov p. , 2001 ,
apj , 549 , 85 ghosh p. , lamb f.k . , 1979 , apj , 232 , 259 ghosh p. , lamb f.k . ,
pethick c.j . , 1977 ,
apj , 217 , 578 jonker p.g . , mndez m. , van der klis m. , 2000 , apj , 540 , 29 jonker p.g . , mndez m. , van der klis m. , 2002 , mnras , 336 , l1 lai d. , 1999 , apj , 524 , 1030 lai d. , 2003 , apj , 591 , l119 lamb f.k .
, miller m.c .
2001 , apj , 554 , 1210 lamb f.k .
, boutloukos s. , van w.s . ,
chamberlain r.t . , lo k.h . ,
clare a. , yu w.f . ,
miller m.c . , 2008 ,
submitted to apj ( astro - ph/0808.4159 ) landau l.d .
, lifshitz e.m . , 1976 ,
mechanics , 3rd edition , course of theoretical physics , vol .
1 , pergamon press , oxford and new york lee w.h . , abramowicz m.a . , kluiak w. , 2004 , apj , 603 , l93 lee u. , 2008 , mnras , 385 , 2069 li x.d . , zhang c.m . , 2005 , apj , 635 , l57 linares m. , van der klis m. , altamirano d. , markwardt c.b .
2005 , apj , 634 , 1250 markwardt c.b .
, strohmayer t.e . , swank j. h. , 1999 , apj , 512 , l125 mndez m. , belloni t. 2007 , mnras , 381 , 790 mndez m. , van der klis m. , 1999 , apj , 517 , l51 mndez m. , van der klis m. , wijnands r. , ford e.c . , van paradijis j. , vaughan b.a . , 1998 ,
apj , 505 , l23 migliari s. , van der klis m. , fender r.p . , 2003 ,
mnras , 345 , l35 miller m.c . , lamb f.k . ,
psaltis d.,1998 , apj , 508 , 791 osherovich v. , titarchuk l. , 1999 , apj , 522 , l113 psaltis d. et al . , 1998 , apj , 501 , l95 rezania v.,samson j.c . , 2005 ,
a&a , 436 , 999 smith d.a . ,
morgan e.h . , bradt h. , 1997 , apj , 479 , l137 stella l. , vietri m. , 1999 , phys .
lett . , 82 , 17 stella l. , vietri m. , morsink s. , 1999 , apj , 524 , l63 strohmayer t.e . , zhang w. , swank j.h . , smale a. , titarchuk l.,day , c. , lee u. , 1996 , apj , 469 , l9 strohmayer t.e . , markwardt c.b . , kuulkers e. , 2008 , apj , 672 , l37-l40 van der klis m. , wijnands r.a.d .
, horne k. , chen w. , 1997 , apj , 481 , l97 van der klis m. , 2006 , in lewin w. h. g. , van der klis m. , eds , compact stellar x - ray sources .
cambridge univ .
press , cambridge , p. 39
van straaten s. , ford e.c . , van der klis m. , mndez m. , kaaret p. , 2000 ,
apj , 540 , 1049 van straaten s. , van der klis m. , di salvo t. , belloni t. , psaltis d. , 2002 , apj , 568 , 912 van straaten s. , van der klis m. , mndez m. , 2003 , apj , 596 , 1155 wijers r. a. m. j. , pringle j. e. 1999 , mnras , 308 , 207 wijnands r.a.d . , van der klis m. , 1997 , apj , 482 , l65 wijnands r.a.d .
, van der klis m. , van paradijs j. , lewin w.h.g . ,
lamb f.k . , vaughan b. , kuulkers e. , 1997 , apj , 479 , l141 wijnands r.a.d . ,
van der klis m. , homan j. , chakrabarty d. , markwardt c.b .
, morgan e.h
. , 2003 , nature , 424 , 44 yin h.x .
, zhang c.m . , zhao y.h .
, lei y.j . , qu j.l . , song l.m . , zhang f. , 2007 , a&a , 471 , 381 zhang c.m . , 2004 , a&a , 423 , 401 zhang f. , qu j. , zhang c.m .
, chen w. , li t.p . , 2006 , apj , 646 , 1116 | we suggest an explanation for the twin kilohertz quasi - periodic oscillations ( khz qpos ) in low - mass x - ray binaries ( lmxbs ) based on magnetohydrodynamics ( mhd ) oscillation modes in neutron star magnetospheres . including the effect of the neutron star spin , we derive several mhd wave modes by solving the dispersion equations , and propose that the coupling of the two resonant mhd modes may lead to the twin khz qpos .
this model naturally relates the upper , lower khz qpo frequencies with the spin frequencies of the neutron stars , and can well account for the measured data of six lmxbs .
[ firstpage ] accretion : accretion discs x - rays : binaries stars : magnetic fields |
it is well known that thermal radiative transfer in the near field is extremely different compared to classical radiative transfer in the far field .
interference effects and , in particular , near field effects due to tunneling of evanescent waves play a vital role . when the objects involved in energy transfer can support surface waves , the heat transfer can be enhanced by orders of magnitude compared to far field values . the near field exchange between two half spaces has been well documented in literature @xcite .
theoretical analysis of heat transfer between a sphere and a plane or between two spheres are limited to the sphere being approximated by a point dipole @xcite . in @xcite , the authors outlined a method capable of dealing with near field radiative transfer between a sphere and a flat substrate when retardation effects can be neglected . however , numerical difficulties prevented them from obtaining a solution to the problem .
it is possible to derive an asymptotic expression for radiative transfer between two large spheres separated by a very small gap ( gap is very small compared to the radius of either of the spheres ) from the well known results of radiative transfer between two semi infinite objects @xcite . this idea is used extensively in determining van der waals or casimir force between macroscopic curved objects and is known as the proximity force approximation @xcite and expressed as the proximity force theorem @xcite . the usage of the proximity force approximation to determine near field radiative transfer between curved surfaces has not been verified by other numerical solutions or by experiments .
generally , the diameter of the sphere involved in the experiments range from a few microns to a few tens of microns , which is no longer in the point dipole approximation .
hence a complete scattering solution to the problem is necessary to verify with experiments as well as gauging the validity of simpler models . in this paper
we investigate , for the first time , the near field radiative heat transfer between two spheres using the dyadic green s function ( dgf ) of the vector helmholtz equation @xcite and the fluctuation
dissipation theorem @xcite .
this formalism of fluctuational electrodynamics was pioneered by rytov @xcite and is used widely for analyzing near field radiative transfer @xcite . though the emphasis of this work is on near field radiative transfer , it should be pointed that this formalism includes far field contribution to radiative transfer .
electromagnetic scattering by a sphere has been very well studied since the seminal work of mie , almost a century ago .
the two sphere scalar and vector scattering problems have also been investigated for almost the same amount of time by many authors @xcite .
the two sphere problem involves expansion of the field in terms of the vector spherical waves of each of the spheres and re - expansion of the vector spherical waves of one sphere in terms of the vector spherical waves of the second sphere in order to satisfy the boundary conditions .
the two sphere problem , and multiple sphere scattering in general , is especially tougher due to the computational demands of determining translation coefficients for the ( vector ) spherical wave functions @xcite .
recurrence relations for the scalar @xcite and vector spherical waves @xcite have reduced the computational complexity considerably . in this work , the dgf for the two sphere configuration is determined by satisfying the boundary conditions for fields on the surface of the two spheres .
the translation coefficients are determined using the recurrence relations in @xcite .
the paper is arranged as follows . in section [ sec : pft ] , simplified , asymptotic results for the radiative heat transfer between two spheres of equal radii , based on the dipole approximation and proximity force approximation are presented . in section [ sec : emag ] the dgf formulation and fluctuation dissipation theorem are introduced .
the poynting vector is expressed in terms of the dgf and material properties . in section [ sec : twosph ] the two sphere problem is described and the dgf for this configuration is determined in terms of the vector spherical waves of the two spheres . in section [ sec : radflux ] , the expression for radiative flux , and thus the spectral conductance , from one sphere to another is determined . details regarding the convergence of the series solution and numerical solutions for sphere sizes up to 20 @xmath0 m in diameter is presented in section [ sec : numres ] .
the purpose of this section is to present an asymptotic results for the radiative heat transfer between two spheres in the dipole limit as well as the proximity force approximation limit .
it is possible to define a radiative conductance between the two spheres at temperatures @xmath1 and @xmath2 as : @xmath3 where @xmath4 ( units wk@xmath5 ) is the radiative conductance , @xmath6 is the rate of heat transfer between the two spheres at @xmath1 and @xmath2 .
it should be noted that @xmath4 is a function of temperature . if the radii of the two spheres involved in near
field radiative transfer are much smaller than the thermal wavelength ( @xmath7 m at 300 k ) , the spheres can be treated as point dipoles and the conductance between the spheres is given by @xcite : @xmath8 where @xmath9 ; @xmath10 and @xmath11 are the imaginary parts of the polarizability of the spheres ; @xmath12 is the center to center distance between the spheres .
this @xmath13 behavior of conductance between two spheres is valid only when @xmath14 and @xmath15 @xcite . calculating the forces or heat transfer between two curved objects as the size of the objects increases become computationally difficult . for spheres of large diameters and much smaller gaps ,
the proximity force approximation is very useful in calculating the forces between the objects and used widely to calculate van der waals or casimir forces @xcite with the knowledge of the same forces between two half spaces .
the proximity force approximation was used to calculate the conductance between a sphere and a flat surface @xcite from the results of the radiative heat transfer between two half
spaces @xcite .
the conductance per unit area for flat plates is the radiative heat transfer coefficient , @xmath16 ( units wm@xmath17k@xmath5 ) .
the heat transfer coefficient @xmath18 in the near field , especially when dominated by surface polaritons , is known to vary as @xmath19 , where @xmath20 is the gap between the half
spaces @xcite .
the spheres are separated by a minimum gap @xmath20 .
we will assume that the spheres are of equal radii , @xmath21 .
the conductance between two spheres is computed by approximating the spheres to be flat surfaces of varying gap . by doing
so we get a relation between @xmath4 and @xmath16 given by @xmath22 since @xmath18 varies as @xmath19 , it is expected that @xmath23 varies as @xmath24 .
a @xmath24 variation of conductance as determined by a more rigorous theory can be taken as evidence of the validity of the proximity type approximation at the value of the gap . despite their origin in fluctuations of the electromagnetic field ,
a significant difference between force and flux is that the force decays to a negligible quantity in the far field whereas thermal flux attains a finite value in the far field .
this implies that those parts of the spheres with larger gaps contribute very little to force whereas they could contribute significanly to flux because of the larger areas involved .
hence a proximity force type approximation would be valid only when the heat transfer is dominated by contributions from the near field region .
therefore , we expect the result of eq .
[ eqn : htpft ] to be valid only when @xmath21 is small enough that near field radiation dominates and @xmath25 .
the discussion in section [ sec : numres ] shows that this is indeed true .
all the materials are assumed to be non magnetic and defined by a complex , frequency dependent dielectric function , @xmath26 . to compute the radiative transfer we follow the method pioneered by rytov @xcite in which the source for radiation is the thermal fluctuations of charges .
the fourier component of the fluctuating electric field , @xmath27 , and magnetic field , @xmath28 , at any point , @xmath29 , outside a volume containing the sources is given by @xcite : @xmath30 where @xmath31 and @xmath32 , the dyadic green s functions due to a point source at @xmath33 , are related by @xmath34 ; @xmath35 is the fourier component of the current due to thermal fluctuations ; and @xmath36 is the permeability of vacuum .
the integration is performed over the entire volume @xmath37 containing the source .
the dgfs themselves obey the following equations @xcite : @xmath38 where @xmath39 is the identity dyad and @xmath40 is the dirac delta function . at the boundary between two dielectric ( possibly lossy ) materials ,
the dgf satisfies the following boundary conditions to ensure continuity of tangential electric and magnetic fields : @xmath41 @xmath42 where @xmath43 and @xmath44 are points on either side of the boundary and @xmath45 is a unit normal to the boundary surface at @xmath43 ( or @xmath44 ) . in order to compute the spectral poynting vector at @xmath29
, we must compute the cross spectral density of @xmath46 and @xmath47 , @xmath48 , where the @xmath49 denotes the complex conjugate , the brackets denote a statistical ensemble average , and @xmath50 and @xmath51 refer to the three cartesian components ( @xmath52 ) . from eq .
( [ eqn : ehgreen ] ) , we can write an expression for @xmath53 as : @xmath54 the fluctuation dissipation theorem states that the cross spectral density of different components of a fluctuating current source in equilibrium at a temperature t is given by @xcite : @xmath55 where @xmath56 is the imaginary part of the dielectric function of the source , @xmath57 is the permittivity of vacuum , and @xmath58 is given by @xmath59 , where @xmath60 is planck s constant and @xmath61 is boltzmann s constant . using eq .
( [ eqn : eihj ] ) and eq .
( [ eqn : fdt ] ) , we have @xmath62 where the superscript @xmath63 stands for the transpose of the dyad . once the green s function for the given configuration is determined , the above integral is computed numerically .
determining the dgf is not a trivial task and the next section is devoted to determining the dgf in the case of the two sphere configuration .
the configuration of the two spheres is shown in fig . [
fig : twosphgeom ] . at the center of each sphere is a coordinate system . without loss of generality ,
the two spheres are arranged such that the z axes of both coordinate systems pass through the line joining the centers .
the x axis ( and y axis ) of both systems are parallel to each other so that a given point in space has the same @xmath64 coordinate value in both systems .
the two spheres are at temperatures @xmath1 and @xmath2 . in order to determine the radiative transfer
, we have to determine the dgf when the source point is in the interior of one of the spheres .
we shall take the dirac
delta source point to be in the interior of sphere a. the most convenient way of dealing with dgf in spherical coordinates is to expand the dgf in terms of vector spherical waves @xcite , which are solutions of @xmath65 the vector spherical waves we will need are given by @xcite : @xmath66 @xmath67 where @xmath68 and @xmath69 are vector spherical waves of order @xmath70 .
@xmath71 can take integer values from @xmath72 to @xmath73 .
for each @xmath71 , @xmath74 .
the superscript @xmath75 refers to the radial behavior of the waves . for @xmath76 ,
the @xmath77 and @xmath78 waves are regular waves and remain finite at the origin and @xmath79 is the spherical bessel function of order @xmath71 . for @xmath80 ,
the @xmath77 and @xmath78 waves are outgoing spherical waves that are singular at the origin and @xmath81 is the spherical hankel function of the first kind of order @xmath71 .
the radial function @xmath82 .@xmath83 , @xmath84 , and @xmath85 are vector spherical harmonics of order @xmath70 .
the three vector spherical harmonics can be expressed in terms of the spherical harmonics , @xmath86 as : two sphere configuration .
two non - overlapping spheres of radii a and b are separated by a distance d. at the center of each sphere is a spherical coordinate system oriented such that the two spheres lie along the common z axis .
the x axes and y axes are also oriented such that for a given location in space , the @xmath64 coordinate is the same in both coordinate systems . in this figure ,
the point p has coordinates @xmath87 and @xmath88 such that @xmath89 .
region a(b ) refers to the interior of sphere a(b ) .
region c is the exterior of both spheres and is taken to be vacuum.,height=226 ] @xmath90 @xmath91 @xmath92 the vector spherical harmonics are orthonormal to each other and satisfy the following relation : @xmath93 where the integration domain @xmath94 refers to the surface of a sphere of unit radius and @xmath95 is a differential area element on such a sphere .
the vector spherical waves @xmath68 and @xmath69 are related by @xmath96 and @xmath97 .
any solution to eq .
[ ref : vechelm ] can be expressed as a linear combination of the vector spherical waves .
further properties of spherical harmonics and vector spherical harmonics useful for this analysis is included in appendix [ sec : appvsw ] and appendix a of @xcite . in this particular case ,
the field is a linear combination of vector spherical waves of the two coordinate systems as shown in fig .
[ fig : twosphgeom ] . to satisfy the boundary conditions on the surface of each sphere the vector spherical waves of one coordinate system should be expressed in terms of the vector spherical waves of the other coordinate system .
this is what is achieved by means of translation addition theorems for vector spherical waves .
the vector translation addition theorem @xcite states that : [ eqn : vectrans ] @xmath98 \end{split}\ ] ] @xmath99 \end{split}\ ] ] the position vectors @xmath100 and @xmath101 refer to the same location in space in coordinate systems a and b respectively .
computing the coefficients @xmath102 and @xmath103 has been the topic of many publications @xcite .
generally , the expressions for the coefficients require calculations of wigner 3j symbols which involve calculations of large number of factorials , making it computationally expensive .
recurrence relations for computing the coefficients efficiently have been proposed by chew @xcite . in the case of the two sphere problem , with translation along the z axis alone ,
[ eqn : vectrans ] simplifies so that the coefficients are non zero for @xmath104 alone . in the region c ( exterior to both spheres )
, the electric and magnetic fields should be expanded in terms of outgoing vector spherical waves of both coordinate systems so that waves decay as @xmath105 as @xmath106 .
hence @xmath107 and @xmath108 need to be expressed in terms of @xmath109 and @xmath110 on the surface of sphere a and vice versa . since @xmath111 for all points on the surface of sphere a , only the regular vector spherical waves or @xmath112 and @xmath113 should be used .
in addition to @xmath114 and @xmath115 , we will also need @xmath116 and @xmath117 , which can be obtained through symmetry relations @xcite . for further details regarding the computation of the recurrence relations , the reader is referred to @xcite .
the dgf for any configuration can be split into two parts - one that corresponds to a dirac delta source in an infinite medium , @xmath118 and one that takes into account the scattering , @xmath119 . in this case , the source point is confined to the interior of sphere a. the dgf for source point in sphere a , assuming the whole space to of the same material , is given by : @xmath120 in particular , we are interested in the case where @xmath121 since the source is inside the sphere a whereas the boundary of interest is the surface of the sphere . the part of the dgf that depends on the boundaries takes different forms in the three regions , a , b , and c. inside a , the dgf is a combination of @xmath118 and @xmath119 , whereas outside a the dgf is entirely @xmath119 . each term , @xmath122 or @xmath123 , in eq .
[ eqn : homodgf ] can be thought of as an independent vector spherical waves that produces scattered waves , i.e. coefficients of scattered waves due to @xmath124 ( or @xmath125 ) are completely decoupled from vector spherical waves of other orders .
let us consider the scattered field due to @xmath124 .
the scattered field in the three regions is given by : @xmath126 & \text { , in a } \\ ik_f \big[\big(c^{lm}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_a ) + c^{ln}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_a)\big ) + \big(d^{lm}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_b ) + d^{ln}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_b)\big ) \big ] & \text { , in c }
\\ ik_b \big[\big(e^{lm}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_b\bm{r}_a ) + e^{ln}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_b\bm{r}_a)\big ) + \big(f^{lm}_{\nu m}\bm{m}_{\nu m}^{(1)}(k_b\bm{r}_b ) + f^{ln}_{\nu m}\bm{n}_{\nu m}^{(1)}(k_b\bm{r}_b)\big ) \big ] & \text { , in b } \end{cases}\ ] ] where the symbol @xmath127 refers to the greater of @xmath128 and 1 .
@xmath129 is a coefficient of a vector spherical waves of order @xmath130 that it is produced by a vector spherical waves of order @xmath70 .
the superscript @xmath131 ( @xmath132 ) is to indicate that it is a coefficient of a @xmath77 ( @xmath78 ) wave . in practice ,
the upper limit for the summation is resticted to a value @xmath133 which depends on @xmath134 .
the appropriate value of @xmath133 will be discussed in section [ sec : numres ] . using eq .
[ eqn : scatfield ] , eq .
[ eqn : elecbounda ] , and eq . [ eqn : elecboundb ] , the following set of coupled linear equations can be obtained for the coefficients of the vector spherical waves in the scattered field in region c : [ eqn : scateqn ] @xmath135 = p^{m}_{\eta}\delta_{\eta l } \end{split}\ ] ] @xmath136 = 0 \end{split}\ ] ] @xmath137 = 0 \end{split}\ ] ] @xmath138 = 0 \end{split}\ ] ] where @xmath139 ranges from @xmath127 to @xmath133 .
@xmath140 and @xmath141 are mie coefficients that one encounters in the scattering of a spherical wave by a single sphere and are given by @xmath142 @xmath143 expressions for @xmath144 and @xmath145 are obtained by replacing @xmath146 and @xmath147 by @xmath148 and @xmath149 respectively . if the original wave is @xmath125 instead of @xmath124 , the only difference is that the right hand side ( rhs ) of eq .
[ eqn : scateqn1 ] becomes 0 and the rhs of eq .
[ eqn : scateqn2 ] becomes @xmath150 .
@xmath151 and @xmath152 are given by @xmath153 @xmath154 + for a given @xmath128 , we have @xmath155 @xmath77 waves and @xmath156 @xmath78 waves .
we see that the left hand side ( lhs ) for a given value of @xmath128 remains the same while the only difference is in the rhs @xmath157 and @xmath158 .
once all the coefficients in eq .
[ eqn : scateqn ] are obtained , the dgf due to scattering , @xmath119 , and its curl , @xmath159 , can be written as : @xmath160 \bm{m}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) + \\ \left [ \begin{matrix } \big(c^{'lm}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_a ) + c^{'ln}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_a)\big ) + \\
\big(d^{'lm}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_b ) + d^{'ln}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_b)\big ) \end{matrix}\right]\bm{n}_{l ,-
m}^{(1)}(k_a\bm{r'}_a ) \end{matrix}\right]\ ] ] @xmath161 \bm{m}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) + \\
\left [ \begin{matrix } \big(c^{'lm}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_a ) + c^{'ln}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_a)\big ) + \\ \big(d^{'lm}_{\nu m}\bm{n}_{\nu m}^{(3)}(k_f\bm{r}_b ) + d^{'ln}_{\nu m}\bm{m}_{\nu m}^{(3)}(k_f\bm{r}_b)\big ) \end{matrix}\right]\bm{n}_{l ,-
m}^{(1)}(k_a\bm{r'}_a ) \end{matrix}\right]\ ] ]
the radiative heat transfer between the two spheres is calculated from the poynting vector normal to the surface of sphere b , which in turn depends on the tangential fields on the surface of sphere b. the expression for the dgf can be modified to reflect tangential and normal fields on the surface of sphere b by using eq .
[ eqn : scateqn3 ] and eq .
[ eqn : scateqn4 ] in eq .
[ eqn : scatdgf ] and eq .
[ eqn : scatdgfmag ] and eliminating @xmath162 and @xmath163 to result in the following equations : @xmath164 \bm{m}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) \\
\\ + \left [ \begin{matrix } \text{similar terms with primed coefficients } \end{matrix}\right]\bm{n}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) \end{matrix}\right ] \end{split}\ ] ] @xmath165 \bm{m}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) \\
\\ + \left [ \begin{matrix } \text{similar terms with primed coefficients } \end{matrix}\right]\bm{n}_{l ,- m}^{(1)}(k_a\bm{r'}_a ) \end{matrix}\right ] \end{split}\ ] ] using eq .
[ eqn : eihjavg ] , eq .
[ eqn : scatdgfsim ] , eq .
[ eqn : scatdgfmagsim ] , eq .
[ eqn : veccross23 ] , eq .
[ eqn : veccross32 ] , eq .
[ eqn : mmsource ] , eq .
[ eqn : nnsource ] , and some algebraic manipulation to yield this expression for the spectral radiative transfer between the two spheres , one at temperature @xmath1 and the other at @xmath2 : @xmath166\im \left(\frac{1}{x_{_l}(a ) } \right ) |x_{_l}(b)|^2 + \\
\left[\im \left(\frac{1}{x_{_\beta}(b ) } \right ) \left|\frac{r^{(1)}_{l}(k_a a ) d_{\beta m}^{'l m}}{z^{(1)}_{\beta}(k_fb ) } \right|^2 - \im \left(\frac{1}{y_{_\beta}(b ) } \right ) \left|\frac{r^{(1)}_{l}(k_fa ) d_{\beta m}^{'l n}}{r^{(1)}_{\beta}(k_fb ) } \right|^2\right]\im \left(\frac{1}{y_{_l}(a ) } \right ) |y_{_l}(b)|^2 \end{matrix}\right ] \end{split}\ ] ] where @xmath167 and @xmath168 . just as in eq .
[ eqn : radcondpft ] , it is possible to define a spectral radiative conductance between the two spheres at a temperature @xmath63 ( @xmath169 ) as : @xmath170\im \left(\frac{1}{x_{_l}(a ) } \right ) |x_{_l}(b)|^2 + \\
\left[\im \left(\frac{1}{x_{_\beta}(b ) } \right ) \left|\frac{r^{(1)}_{l}(k_a a ) d_{\beta m}^{'l
m}}{z^{(1)}_{\beta}(k_fb ) } \right|^2 - \im \left(\frac{1}{y_{_\beta}(b ) } \right ) \left|\frac{r^{(1)}_{l}(k_fa ) d_{\beta m}^{'l n}}{r^{(1)}_{\beta}(k_fb ) } \right|^2\right]\im \left(\frac{1}{y_{_l}(a ) } \right ) |y_{_l}(b)|^2 \end{matrix}\right ] \end{split}\ ] ] where @xmath171 .
it can be seen from eq .
[ eqn : speccond ] that the spectral conductance has units of @xmath172 and can split into two parts , one that depends on temperature and the other that is obtained from the dgf of the two sphere problem .
the radiative conductance between the two particles that one would measure in an experiment is the integral of @xmath173 .
to ensure that the program written to determine the near field radiative transfer heat transfer is not misbehaving , a few tests can be performed .
one of them is agreement between the numerical results and the analytic expression for conductance in the point dipole limit . for spherical particles in the 1 nm to 50 nm radius ,
the numerical results agree well with the expression for conductivity in the point dipole limit @xcite .
in addition to this another test to ensure the correctness of the method is based on the principle of detailed balance .
the radiative conductance between two spheres of arbitrary radii must be independent of the numbering scheme for naming the particles , i.e @xmath175 , where the subscripts 1 and 2 refer to the two spheres .
this is necessary to ensure that when the two spheres are at the same temperature the net heat transfer between the two spheres is zero .
it is indeed seen from results shown in table [ table : conductancerecip ] that by switching the position if the spheres , keeping the gap the same , results in the same value of conductance ( the relative errors are generally of the order 10@xmath176 ) .[table : conductancerecip]conductance obtained for spheres of unequal radii . by swapping the radii of the spheres , it is seen that the value of conductance remains the same .
[ cols="^,^,^,^",options="header " , ] though the two - sphere scattering problem has been discussed in literature , the near
field interaction between the two spheres has not been analyzed in detail .
in particular , the number of terms required for convergence for the near field problem has not been mentioned at all . for mie scattering by a single sphere of radius @xmath147 and wavevector magnitude @xmath177 the number of terms for convergence , @xmath178 , is given by @xcite : @xmath179 for the two sphere problem a slightly different criterion based on the center to center distance between the two spheres is proposed and given by @xcite : @xmath180 where @xmath181 is the base of the natural logarithm . both eq .
[ eqn : mieconvff ] and eq .
[ eqn : twosphff ] are valid criteria for computing far field quatities , like the scattering coefficient . for the near field problem
, we expect the gap between the two spheres to be of great importance . to determine the number of terms required for convergence of near field quantities we seek parallels to the much simpler and often investigated problem of near field transfer between two half spaces . in the near field two half space problem , the equivalent of the number of terms for convergence is the truncation for the in plane wave vector . for a gap @xmath20 between the two half spaces ,
the predominant contribution to radiative flux in those frequency regions where electromagnetic surface waves are important is from in plane wave vectors up to the order of @xmath182 .
this is true when @xmath183 , where @xmath184 is the dielectric function of the material of the sphere .
this dielectric function of silica between the frequencies of 0.04 ev to 0.16 ev satisfies this condition , in which range the maximum value of @xmath185 is @xmath186 . if we can draw an analogy between the in plane wave vectors in the two half space problem and the two sphere problem , we can propose a criterion for convergence . in the case of the two - sphere problem ,
the in plane wave vector equivalent is given by the wavelength of periodic variations on the surface of the sphere .
a given vector spherical harmonic , @xmath187 , @xmath188 corresponds to a variation @xmath189 along the equator of the sphere .
the period of this particular vector spherical harmonic ( along the equator ) is @xmath190 and the corresponding wave vector is @xmath191 , where @xmath147 is the radius of the sphere . for a given gap @xmath20 between two spheres of radii @xmath147 ( equal for now ) , the number of terms @xmath178 for convergence should be chosen such that @xmath192 .
hence the convergence criterion for near field effects is given by : @xmath193 for spheres of unequal radii the number of terms depends on the greater of the two radii .
the convergence criterion proposed here is an upper limit and depending on the optical properties of the spheres , it could be considerably lesser . for spheres which exhibit surface wave resonances , eq . [ eqn : twosphnf ] is a valid convergence criterion as shown by our numerical investigations . depending on the configuration of the spheres ,
the criterion for convergence is given by @xmath194 .
since silica spheres are easily available for experimental investigation , we will present our numerical results of radiative transfer between two silica spheres of equal radii at 300 k. in addition , silica is a polar material and hence supports surface phonon polaritons in certain frequency ranges .
the dielectric function of silica is taken from @xcite .
the real part of the dielectric function is negative for silica in two frequency ranges in the ir from 0.055 to 0.07 ev and 0.14 to 0.16 ev .
it is expected that ( and shown later ) that surface phonon polariton resonances occur in the these frequency ranges .
to confirm the prediction of eq .
[ eqn : twosphnf ] , the contribution to the spectral conductance from each @xmath70 mode is plotted as a function of @xmath71 for different values of @xmath128 in fig .
[ fig : nmaxconvres ] .
the spectral conductance is plotted at two frequencies 0.061 ev and 0.045 ev .
the surface polariton mode exists at 0.061 ev but not at 0.045 ev .
plotted in this figure is the contribution to spectral conductance as a function of @xmath71 for @xmath195 at 0.061 ev and 0.045 ev for two spheres of equal radii @xmath196 @xmath0 m at a gap of 100 nm ( @xmath197 @xmath0 m ) .
surface phonon polaritons contribute significantly to the radiative transfer at 0.061 ev and not at 0.045 ev .
hence the number of terms required for convergence is significantly lesser than that prescribed by eq .
[ eqn : twosphnf],height=226 ] in fig .
[ fig:3 ] the spectral conductance at 0.061 ev for @xmath128 = 0 between two spheres of radius 10 @xmath0 m is plotted as a function of @xmath71 for two different gaps 100 nm and 200 nm . as the gap doubles , the number of terms required for convergence approximately reduces by half , confirming eq .
[ eqn : twosphnf ] .
for instance , the spectral conductance is @xmath198 at @xmath199 for 100 nm gap and @xmath200 for 200 nm gap .
the contribution to spectral conductance from smaller values of @xmath71 ( @xmath201 ) , corresponding to propagating waves , does not change much with variation in gap .
it is the contribution from larger values of @xmath71 , corresponding to surface waves , that changes appreciably with gap
. based on these results we use at least @xmath202 terms in our computations . because of the computation difficulties in solving eq .
[ eqn : scateqn ] , we present results for a maximum radius to gap ratio of 100 for computations at one frequency and 100 frequency points ( unequally spaced , with greater density in the resonant parts of the spectrum ) over the spectrum from 0.04 ev to 0.16 ev . equation [ eqn : scateqn ] is solved using the software package mathematica . plot of spectral conductance at 0.061 ev between two spheres with equal radii of 10 @xmath0 m at gaps of 100 nm and 200 nm .
the curves plotted are for m = 0 . , height=226 ] as mentioned earlier , the symmetry associated with the two - sphere problem allows for solving for the contribution from each value of the @xmath128 independently , starting from @xmath203 and proceeding with increasing values of @xmath128 . as @xmath128 increases , the contribution to conductance decreases as shown in fig . [
fig : convvsm ] .
the computation is stopped when a vacule of @xmath128 is reached such that the contribution to conductance is less than 5 @xmath204 10@xmath205 times the contribution from @xmath203 .
even though the contribution to conductance is significant for terms with @xmath206 , the contributions from @xmath128 drops much faster .
this is fortunate - the time taken to compute the conductance is proportional to the number of values of @xmath128 required .
+ unlike the case of near field radiative transfer between two half spaces , where the conductance is a function of only the gap between the half spaces ( and the optical properties of the half spaces and interveing medium ) , the conductance in the case of sphere to sphere radiative transfer varies as a function of the gap as well as the size of the sphere .
a gradual transition occurs from a region of near field dominated radiation to that of far field dominated radiation . in fig .
[ fig : spectralcond1mic ] the spectral conductance between two silica spheres of 1 @xmath0 m radius is plotted as a function of frequency .
we see from the two peaks that the conductance is dominated by the surface phonon polariton regions . the spectral conductance between the two spheres for larger radii , shown in fig .
[ fig : spectralcond5mic ] , displays several features of interest .
the ratio of radius to gap is maintained the same as in fig .
[ fig : spectralcond1mic ] .
though the height of the peaks remain approximately the same in this figure as well as fig .
[ fig : spectralcond1mic ] , the significant difference is from the contribution from those frequency regions that do not support surface polaritons ( 0.040.055 ev , 0.070.14 ev ) .
the contribution to the conductance from these ranges do not vary significantly with gap ( as long as gap @xmath207 radius ) .
the relation between the ratio of gap to radius and the contribution to spectral conductance is also illustrated in fig .
[ fig : speccondsamedbyrratio ] . in fig .
[ fig : speccondsamedbyrratio ] , the spectral conductance of spheres of radii 1 @xmath0 m , 2 @xmath0 m , and 5 @xmath0 m at gaps of 100 nm , 200 nm , and 500 nm respectively are shown .
since the ratio of gap to radius is a constant ( 0.1 ) , we expect from the asymptotic theory that value of spectral conductance should also be the same in all three cases .
we see from the data that this is approximately the case in the regions where electromagnetic surface waves dominate the heat transfer . in the rest of the region , where near field radiative transfer is non resonant
, increasing radius leads to increased contribution from propagating waves .
the spectral conductance as the radius of the spheres is increased to 20 @xmath0 m is shown in fig .
[ fig : spectralcond20mic ] .
the increased contribution from the non resonant parts of the spectrum is evident from the graph .
this has an important implication from an experimental point of view
. we should be careful not to increase the size of the sphere to such an extent that the resonant radiative transfer is swamped by the non resonant radiative transfer .
for a 20 @xmath0 m sphere , we can see that the increase in conductance as the gap decreases is still predominantly due to electromagnetic surface waves as shown in fig . [
fig : spectralcond20micdiff ] . in fig .
[ fig : spectralcond20micdiff ] , the increase in spectral conductance as the gap is decreased from 2000 nm is plotted . compared to fig .
[ fig : spectralcond20mic ] , the signature of electromagnetic surface waves is clearer from this plot . in fig . [
fig : speccondat61mev ] the spectral conductance at 0.061 ev ( corresponding to the first peak in fig .
[ fig : spectralcond1mic ] ) between two spheres is plotted as a function of gap for different values of the radii .
the exponent of a power law fit ( of the form @xmath208 ) to the data points in fig .
[ fig : speccondat61mev ] is -1.001 , -0.984 , -0.9024 , -0.7781,-0.6119 for radius 1 @xmath0 m , 4 @xmath0 m , 10 @xmath0 m , 20 @xmath0 m , and 40 @xmath0 m respectively . as the radius increases from 1 @xmath0 m to 40 @xmath0 m
, the slope of the curve decreases , indicating an increased contribution from propagating waves .
the behavior at smaller radii can be predicted from the variation of conductance with gap between two planes and the proximity force approximation , as discussed in section [ sec : pft ] . for larger diameters ,
the proximity force type approximation is seen to fail .
the frequency limits for the calculation of conductance is taken to be 0.041 ev to 0.164 ev .
the contribution to conductance from the rest of the frequency spectrum at 300 k is not significant for spheres of smaller radii .
however , it is seen from fig .
[ fig : spectralcond20mic ] that frequencies below 0.04 ev contribute to the non resonant heat transfer .
this will not affect the increase in radiative transfer as the gap is decreased because that increase comes predominantly from the regions supporting surface waves .
the total conductance for spheres up to radius of 5 @xmath0 m are plotted against gap in fig .
[ fig : totcond5mic ] .
the slope of -6 ( approximate ) for spheres of radius 20 nm and 40 nm , which can be approximated as point dipoles when the gap between the spheres is much larger than the radius , is in correspondence with the results from the dipole approximation .
however , the dipole theory predicts that the conductance should flatten and reach a finite value as the gap decreases to zero , i.e. , a slope of zero .
what happens in fact is that the near field effects begin contributing as the gap decreases and the slope in fact decreases from -6 to approximately -1 , once again corresponding to the asymptotic theory .
however , as the radius increases to 5 @xmath0 m , the slope decreases further .
this decrease is because of increased contribution from non resonant regions of the spectrum , as seen in fig .
[ fig : spectralcond5mic ] . for larger values of spheres ,
the conductance is plotted in fig .
[ fig : totcondlargedia ] . on a log - log scale as plotted in fig .
[ fig : totcondlargedia ] , the near field effects are apparent for the spheres of 1 @xmath0 m and 2 @xmath0 m .
the reason it does not seem so for the spheres of larger diameter is because of the large contribution from the non resonant parts of the spectrum , as evidenced from the curve corresponding to the results of classical radiative transfer for the 20 @xmath0 m sphere . to understand the effects of near field transfer for larger spheres , the total conductance for 20 @xmath0 m and 25 @xmath0 m spheres
is plotted in fig .
[ fig : totcond10mic20mic ] .
the data points are fit with a curve @xmath209 , where @xmath20 is the gap , @xmath210 is the near field contribution , @xmath211 is the contribution due to non resonant parts of the spectrum ( as well as any changes from classical effects due to increase in view factor ) .
we see that the value of the exponent @xmath212 is 0.5574 for the 20 @xmath0 m sphere and 0.5035 for the 25 @xmath0 m sphere . if the asymptotic theory we valid at these values of the gap , one would have expected a value of @xmath213 .
it is expected that as the gap decreases , the conductance will approach the form predicted in eq .
[ eqn : radcondpft ] by the asymptotic theory ( not confirmed in this paper due to computational restrictions ) .
unlike the two half space problem @xcite , the two sphere problem , when other relavant length scales such as the skin depth are unimportant or @xmath214 ( as in the case of silica spheres ) , where @xmath184 is the dielectric function of the sphere , has three length scales - the wavelength in consideration @xmath215 , the radius of the spheres @xmath147 , and the gap between the spheres , @xmath20 .
depending on the ratio of the length scales , different approximate theories can be used to predict the conductances .
these regions in which different theories are applicable can be represented on a regime map with the two axes representing two non dimensional length scales , as shown in fig . [ fig : regimemap ] and discussed below . 1 . @xmath216 and @xmath217 : in this case classical radiative transfer can be employed
. however , while near field effects may not be important , interference effects can become important .
when @xmath218 , diffraction effects prevent the usage of classical radiative transfer for even emission from a single sphere .
2 . @xmath219 and @xmath220 : when the dipole moment of the particles is the dominant contributor to the radiative transfer , the point dipole approximation can be used .
however , it should be mentioned that even though the gaps are numerically small , this is not a near field effect in the sense discussed in this paper .
@xmath221 : in this case , the conductance is sees to vary linearly with @xmath24 and is indicative of the validity of the proximity approximation in this regime .
the numerical solution to the two sphere problem as discussed in this paper is valid in all regions of the map , limited only by computational demands .
near - field radiative transfer between two spheres has been investigated for the very first time using fluctuational electrodynamics formalism .
numerical results for spheres of equal radii have been presented and analyzed .
it is seen that the proximity force approximation type theory for near field radiative transfer is valid only when the radius and gap satisfy the condition @xmath222 . for larger spheres ,
the proximity force approximation , which is widely used in calculating forces , is not valid for the gaps analyzed in this paper .
the purpose of solving the two sphere problem is to extend the theoretical and numerical formulations of near field radiative transfer to configurations of objects which can be tested experimentally .
the solution to the two - sphere problem gives us an estimate of the value of radiative conductance one can expect from such an experiment . for microspheres with radii of 25 @xmath0 m
, we expect a radiative conductance between two silica spheres around 4.5 nwk@xmath5 at a gap of 200 nm . from fig .
[ fig : totcond10mic20mic ] , we see that we can expect a conductance of 5.6 nwk@xmath5 at a gap of 100 nm .
if the experimental configuration is not a two - sphere configuration but a sphere adjacent to a flat plate , the results of this chapter can be used as a guide to estimating the conductance .
we expect trends to be similar that is we expect the increase in conductance to be of the form @xmath223 , where @xmath20 is the gap between the sphere and the flat plate .
most importantly , we expect @xmath212 to be a number between 0 and 1
. the work is supported by an onr grant ( grant no .
n00014 - 03 - 1 - 0835 ) through university of california , berkeley . | radiative energy transfer between closely spaced bodies is known to be significantly larger than that predicted by classical radiative transfer because of tunneling due to evanescent waves .
theoretical analysis of near field radiative transfer is mainly restricted to radiative transfer between two half spaces or spheres treated in the dipole approximation ( very small sphere ) or proximity force approximation ( radius of sphere much greater than the gap ) .
sphere sphere or sphere plane configurations beyond the dipole approximation or proximity force approximation have not been attempted . in this work
, the radiative energy transfer between two adjacent non overlapping spheres of arbitrary diameters and gaps is analyzed numerically . for spheres of small diameter ( compared to the wavelength ) ,
the results coincide with the dipole approximation .
we see that the proximity force approximation is not valid for spheres with diameters much larger than the gap , even though this approximation is well established for calculating forces . from the numerical results ,
a regime map is constructed based on two non dimensional length scales for the validity of different approximations . |
supermassive black holes are believed to power quasars and other types of agn .
the agn activity manifests itself as radiation from the region around the black hole and as highly collimated outflows from the immediate vicinity of the black hole ( e.g. rees 1984 ) .
two defining properties of astrophysical black holes that can be measured in principle are the black hole mass and spin .
a significant amount of progress has been made toward measuring the masses of supermassive black holes ( e.g. kormendy & richstone 1995 ; ferrarese & ford 2005 ) , though measuring the spin has been more challenging .
general theoretical studies of black hole spins indicate that accretion of gas is likely to produce rapidly spinning black holes ( volonteri et al .
2005 ) , the merger of two nonspinning black holes of comparable mass is likely to produce a rapidly spinning black hole ( gammie , shapiro , & mckinney 2004 ) , and the capture of a smaller companion may cause the hole to spin down ( hughes & blandford 2003 ) . to date , there are only a few direct observations that allow black hole spins to be studied .
observations of seyfert galaxies suggest rapidly rotating black holes in these systems ( wilms et al .
2001 ; fabian et al .
a large spin is indicated by observations of the galactic center black hole ( genzel et al .
2003 ; ashenbach et al .
2004 ) . and
, x - ray observations of active galaxies suggest rapidly rotating black holes in these systems ( crummy et al .
2006 ) .
theoretical predictions for the spins of black holes for fri and frii radio sources ( defined by fanaroff & riley 1974 ) have been made by meier ( 1999 ) in the context of a rotating black hole model , and several predictions of this model are very similar to those obtained in the model of blandford & znajek ( 1977 ) . in this work , a new method of measuring the spin of a supermassive black hole
is proposed .
the method may be applied to sources with large - scale outflows that power radio sources , and , perhaps , other types of sources .
a double radio source is powered by two oppositely directed highly collimated outflows from the vicinity of a supermassive black hole .
the spin energy that is available to power the outflow can be written in terms of the black hole mass and spin . thus ,
if the spin energy and black hole mass can be estimated or constrained , the black hole spin can be estimated or constrained . the method to obtain a measure of the black hole spin
is described in section 2 .
this method is independent of the specific model of energy extraction .
the results obtained by applying the method are presented in section 3 .
a discussion follows in section 4 .
the results are obtained assuming a standard cosmological model with @xmath3 km / s / mpc , @xmath4 , @xmath5 , and zero space curvature .
the energy @xmath6 that can be extracted from a black hole is related to the black hole spin @xmath7 and mass @xmath8 , as discussed , for example , by blandford ( 1990 ) , where @xmath9 is defined in terms of the spin angular momentum @xmath10 , the speed of light @xmath11 and the black hole mass @xmath8 , @xmath12 , and @xmath13 is the gravitational radius of the black hole , @xmath14 .
the spin energy is @xmath15 , where @xmath16 is the irreducible mass of the hole . setting @xmath6 equal to the spin energy and solving for the black hole spin , we find that @xmath17 where @xmath18 , or @xmath19 for @xmath20 .
thus , for sources with outflows powered by the spin energy of a black hole , the black hole spin @xmath1 can be measured when the total energy @xmath6 of the outflow and the black hole mass @xmath8 are known .
determinations of the black hole spin obtained using equation ( 1 ) do not depend upon the details of the energy extraction .
the use of this equation assumes only that the ultimate source of the energy is the black hole spin , and that , during the short period of spin energy extraction , the black hole spin changes only because of the energy extraction .
the ratio @xmath0 provides an important diagnostic of the black hole system .
studies of the ratio @xmath0 provide insight into the physical state of the system at the time the outflow is generated irrespective of whether the outflow is powered by the spin energy of the hole .
the spin energy of the black hole is thought to power the large - scale outflows from agn ( e.g. blandford & znajek 1977 ; rees 1984 ; blandford 1990 ; wilson & colbert 1995 ; daly 1995 ; meier 1999 , 2002 ) .
the total energy @xmath6 that will be channeled through a large - scale outflow can be determined for the most powerful classical double radio galaxies ( e.g. daly & guerra 2002 ; wan , daly , & guerra 2002 ; daly et al .
2008 ; odea et al .
this empirically determined total energy @xmath6 can then be combined with the black hole mass to obtain the ratio @xmath0 and a measure of or bound on the spin parameter of sources using eq .
( 1 ) . if only part of the spin energy is tapped , or if some of the tapped energy is dissipated or radiated away and does not reach the extremities of the radio source , the value of @xmath1 obtained using eq . ( 1 ) will be a lower bound .
if the outflows are powered by energy from some other source , such as an accretion disk , then the energy per unit black hole mass , @xmath0 , could be used to probe the physics of the disk and , perhaps , place a limit on the black hole spin . for models in which the large scale outflow is powered by processes in the accretion disk , the black hole spin may be studied by combining the beam power of the outflow and mass of the black hole .
there are 19 powerful radio galaxies for which we have estimates of both the black hole mass @xmath8 and the total energy @xmath6 .
all of these very powerful frii radio sources have radio powers at least a factor of ten above the classical fri / frii transition , and are referred to as friib sources .
the sources and their properties are listed in table 1 .
the source redshifts range from 0.056 to 1.79 , and the hot spot to hot spot source sizes range from about 100 kpc to over 600 kpc ( odea et al .
. the black hole mass for cygnus a ( 3c 405 ) is obtained from tadhunter et al .
( 2003 ) ; those for the next four sources are obtained from mclure et al .
( 2004 ) , and those for the remaining 14 sources are obtained using equation ( 3 ) of mclure et al .
( 2006 ) where the values of @xmath21 were provided by mclure ( mclure , private communication 2008 ) .
the total energies are obtained from wan , daly , & guerra ( 2000 ) , guerra , daly , & wan ( 2000 ) , and odea et al .
( 2008 ) and converted to the cosmological model adopted here .
the energy from the two sides of each source are combined to obtain the total outflow energy @xmath6 , which is taken to be twice the weighted mean of the outflow energies from each side of the source .
deviations of the magnetic field strength of the extended radio source from the minimum energy value only enter @xmath6 through the normalization of this quantity , as discussed in detail by odea et al .
( 2008 ) . the lobes of cygnus a are used to normalize the sources , and an offset of 0.25 of magnetic field strength from the minimum energy value , appropriate for cygnus a ( carilli et al . 1991 ; wellman , daly , & wan 1997 ) has been adopted .
note that if there is no offset from minimum energy conditions in cygnus a , the empirically determined values of @xmath6 and @xmath0 decreases by about a factor of 5 ( odea et al . 2008 ) , so the values of @xmath1 decrease by about @xmath22 from those discussed below .
the dimensionless ratio of outflow energy to black hole mass , @xmath23 , is obtained for each source and substituted into eq .
( 1 ) to obtain the spin parameter @xmath1 , and the values are listed in table 1 .
the sources have very similar values of @xmath0 and @xmath1 , with no dependence of either parameter on source size or redshift .
the weighted mean values of @xmath0 and @xmath1 for these 19 sources are @xmath24 and @xmath2 , respectively .
interestingly , within the measurement error , the values of @xmath0 and @xmath1 for each source are consistent with the mean value of the 19 sources .
the values of @xmath1 are shown as a histogram in figure 1 and as a function of black hole mass in figure 2 .
figure 3 shows the total outflow energy as a function of black hole mass .
it would appear that the values of @xmath0 and @xmath1 for these friib sources is constant , suggesting that the physical state of the system at the time the outflow is generated is very similar in each of the sources studied .
this could indicate that the outflow is triggered when a particular threshold , given by @xmath25 , is reached . a model that includes a threshold of this type
is discussed by meier ( 1999 ) .
a threshold for the onset of the outflow is also indicated by a comparison of individual source properties with the properties of the full population of sources ( e.g. daly & guerra 2002 ; daly et al .
2008 ) .
there are now many examples of galaxy clusters in which the intracluster medium ( icm ) has been displaced as the result of powerful outflows from agn ( e.g. mcnamara & nulsen 2007 ; rafferty et al .
the displacement of the icm provides a measure of the total energy that has been pumped into the icm by the large - scale outflow from the agn associated with the central dominant galaxy ( cdg ) .
if the highly collimated outflow is powered by the spin energy of the agn , then the energy of the outflow and the mass of the black hole can be combined to solve for the spin @xmath1 of the black hole .
of the 33 galaxy clusters studied by rafferty et al .
( 2006 ) , 29 have estimates of both the mass of the black hole associated with the large - scale outflow and the total energy input to the icm by the agn .
one source , a1068 , does not have an estimate of the total energy input to the icm by the agn , and three sources , rbs 797 , macs j1423.8 + 2404 , and 3c401 , do not have estimates of the black hole mass .
the remaining 29 sources are considered here .
almost all of these sources have fri radio source structure or have amorphous radio structure , with a few exceptions such as cygnus a ( birzan et al . 2008 ) .
equation ( 1 ) was used to estimate the spin of each of the 29 sources for which there is a measure of both the energy @xmath6 and the black hole mass @xmath8 .
the energy @xmath26 and black hole mass for each source are obtained from rafferty et al .
( 2006 ) , where the values of @xmath27 are used since they are obtained in a manner that is very similar to the black hole mass determinations described in section [ frii ] .
the values of @xmath27 listed in table 3 of rafferty et al .
( 2006 ) had been modified by those authors by a factor of 0.35 , and this adjustment has been removed .
this brings the mass estimates for cygnus a ( 3c 405 ) and m84 into good agreement with the independent determinations by tadhunter et al .
( 2003 ) and maciejewski & binney ( 2001 ) , respectively . for simplicity
, the average value of asymmetric error bars was used .
the values obtained for the ratio @xmath0 and spin @xmath1 of each source are listed in table 2 .
the median value of @xmath1 is 0.026 , the average value of @xmath28 is @xmath29 , and the values of @xmath1 range from about 0.001 to 0.4 .
a histogram of the values of @xmath30 is shown in figure 1 , and @xmath1 is shown as a function of black hole mass in figure 2 .
figure 3 shows the total outflow energy as a function of black hole mass . clearly , there is a much broader range of @xmath6 , @xmath0 , and @xmath1 for the cdgs relative to those of powerful classical double radio galaxies .
most of the values of @xmath30 lie in the range from @xmath31 to @xmath32 .
the values of the energy extracted for the 29 cdgs studied include the energy extracted up to the time of observation , and each source is being observed at a different stage in its lifetime .
in addition , relativistic plasma left from a prior outflow event could cause the outflow energy from the current outflow event to be overestimated .
these effects can cause the observed distributions of @xmath0 and @xmath1 to be quite broad .
the range of values of @xmath0 and @xmath1 for the friib radio galaxies studied is remarkably small , while the range of values of @xmath0 and @xmath1 for outflows associated with cdgs is quite broad .
there are two key distinctions between these samples that may partially explain these differences .
firstly , the outflow energy measured for a friib source is an estimate of the total outflow energy that will be produced over the entire lifetime of the outflow , whereas the outflow energy measured for the cdgs is an estimate of the outflow energy produced by the source up to the time of the observation .
catching each cdg outflow at a different phase in the course of its lifetime will cause the distribution of outflow energies for cdgs to be broad ; this is not a factor for the friib sources .
in addition , a prior outflow event from a cdg could cause the outflow energy of a given event to be overestimated .
secondly , the friib radio galaxies are selected from the flux limited 3crr catalogue ( laing , riley , & longair 1983 ) , so only the most powerful sources are detected at each redshift .
interestingly , the total outflow energy @xmath6 measured for powerful 3crr radio galaxies increases as @xmath33 ( odea et al .
2008 ) , while the black hole mass rises as @xmath34 ( mclure et al .
2006 ) , so the ratio @xmath0 is independent of redshift .
the fact that lower mass and lower energy black holes are not observed at high redshift is due to the flux limit of the survey .
the fact that the high mass and high energy sources friib sources are not seen at low redshift may suggest that the higher mass sources are more active at high redshift , or that changes in the environment cause the radio power of the sources to fall below the flux limit of the survey , that is , the sources may be lower power frii sources or fri sources , though they could still have large outflow energies ( e.g. hardee et al . 1992 ; odonoghue , eilek , & owen 1993 ; bicknell 1995 ) .
a selection bias also seems to be present in the cdg sample ; the 6 sources with outflow energies less than @xmath35 are all at very low redshift , and the outflow energy , black hole mass , and black hole spin all tend to increase with redshift . very powerful 3crr radio galaxies are likely to evolve into the cdgs ( lilly & longair 1984 ; best et al . 1998 ; mclure et al . 2004 ) .
thus , the two samples studied are representative of the early and late phases of evolution of these systems .
the systems have black holes with similar mass .
clearly , the friib sources have a very small range of outflow energy per unit black hole mass , @xmath0 .
it would appear that there is a larger range of values for this parameter for cdgs , but , given the factors affecting the measurements of outflow energy from black holes of cdgs , it is possible that these sources have a similarly small intrinsic range of @xmath0 and @xmath1 .
the results obtained here are consistent with the idea that the different radio structures of fri and frii sources are due to differences in their gaseous environments ( e.g. burns et al .
1994 ; bicknell 1995 ) , and suggests that the gaseous environments of these sources has evolved significantly .
the evolution of the gaseous environment could be due in part to the large - scale outflows ( e.g. silk & rees 1998 ; eilek & owen 2002 ) .
it is remarkable that so few sources have values of @xmath1 close to the theoretical limit of unity ; that is , almost all sources have values of @xmath0 much less than the theoretical limit of 0.29 .
all of the friib sources have @xmath36 .
this suggests that each system is in a similar physical state at the time the outflow is generated , and may indicate that the outflow is triggered when a particular threshold is reached .
if the sources have intrinsic values of @xmath1 that start out close to one , each outflow event is tapping a very small fraction of the available spin energy , and each source may have multiple outflow events . the spin values obtained here could be compared with independent measures of the black hole spin .
this may provide a diagnostic of whether it is the black hole spin or the accretion disk that is powering large - scale outflows from these sources .
in addition , the method of using outflow energies and black hole masses to estimate black hole spins presented here may be applicable to other systems . it is a pleasure to thank george djorgovski , brian mcnamara , david meier , and the referee of this paper for very helpful comments and suggestions on this work .
it is also a pleasure to thank ross mclure for providing stellar masses for the fourteen highest redshift sources listed in table 1 .
this work was supported in part by u. s. national science foundation grants ast-0507465 .
lclcll 3c 405 & @xmath37&@xmath38&@xmath39&@xmath40&@xmath41 + 3c 244.1 & @xmath42&@xmath43&@xmath44&@xmath45&@xmath46 + 3c 172 & @xmath47&@xmath48&@xmath49&@xmath50&@xmath51 + 3c 330 & @xmath52&@xmath53&@xmath54&@xmath55&@xmath56 + 3c 427.1 & @xmath57&@xmath58&@xmath59&@xmath60&@xmath46 + 3c 337 & @xmath61&@xmath62&@xmath63&@xmath64&@xmath65 + 3c34 & @xmath66&@xmath67&@xmath68&@xmath69&@xmath65 + 3c441 & @xmath70&@xmath71&@xmath72&@xmath73&@xmath65 + 3c 55 & @xmath74&@xmath75&@xmath76&@xmath77&@xmath78 + 3c 247 & @xmath79&@xmath80&@xmath81&@xmath82&@xmath83 + 3c 289 & @xmath84&@xmath53&@xmath85&@xmath86&@xmath87 + 3c 280 & @xmath88&@xmath89&@xmath85&@xmath90&@xmath83 + 3c 356 & @xmath91&@xmath92&@xmath93&@xmath94&@xmath95 + 3c 267 & @xmath96&@xmath97&@xmath98&@xmath99&@xmath95 + 3c 324 & @xmath100&@xmath101&@xmath102&@xmath103&@xmath87 + 3c 437 & @xmath104&@xmath105&@xmath106&@xmath107&@xmath108 + 3c 68.2 & @xmath109&@xmath110&@xmath111&@xmath112&@xmath113
+ 3c 322 & @xmath114&@xmath115&@xmath116&@xmath117&@xmath118 + 3c 239 & @xmath119&@xmath120&@xmath121&@xmath122&@xmath123 + [ table ] lccccc ms 0735.6 + 7421 & 0.216 & @xmath124&@xmath125&@xmath126&@xmath127 + zw 2701 & 0.214 & @xmath128&@xmath129&@xmath130&@xmath131 + hydra a & 0.055 & @xmath132&@xmath133&@xmath134&@xmath135 + zw 3146 & 0.291 & @xmath136&@xmath137&@xmath138&@xmath139 + mkw 3s & 0.045 & @xmath140&@xmath141&@xmath142&@xmath143 + cygnus a & 0.056 & @xmath144&@xmath145&@xmath146&@xmath147 + pks 0745 - 191 & 0.103 & @xmath148&@xmath149&@xmath150&@xmath151 + hercules a & 0.154 & @xmath152&@xmath125&@xmath153&@xmath154 + sersic 159/03 & 0.058 & @xmath155&@xmath129&@xmath156&@xmath157 + a133 & 0.060 & @xmath158&@xmath159&@xmath160&@xmath161 + perseus & 0.018 & @xmath162&@xmath163&@xmath164&@xmath165 + a1835 & 0.253 & @xmath166&@xmath167&@xmath168&@xmath169 + 4c
55.16 & 0.242 & @xmath170&@xmath171&@xmath172&@xmath173 + a2597 & 0.085 & @xmath174&@xmath141&@xmath175&@xmath176 + a2199 & 0.030 & @xmath177&@xmath178&@xmath179&@xmath180 + 3c 388 & 0.092 & @xmath181&@xmath163&@xmath182&@xmath183 + a1795 & 0.063 & @xmath184&@xmath185&@xmath186&@xmath187 + a4059 & 0.048 & @xmath188&@xmath145&@xmath189&@xmath190 + a2052 & 0.035 & @xmath191&@xmath163&@xmath192&@xmath193 + a2029 & 0.077 & @xmath194&@xmath195&@xmath196&@xmath197 + 2a 0335 + 096 & 0.035 & @xmath198&@xmath171&@xmath199&@xmath200 + a478 & 0.081 & @xmath201&@xmath202&@xmath203&@xmath204 + a85 & 0.055 & @xmath205&@xmath145&@xmath206&@xmath207 + pks 1404 - 267 & 0.022 & @xmath208&@xmath209&@xmath210&@xmath211 + a262 & 0.016 & @xmath212&@xmath141&@xmath213&@xmath214 + hcg 62 & 0.014 & @xmath215&@xmath209&@xmath216&@xmath217 + centaurus & 0.011 & @xmath218&@xmath141&@xmath219&@xmath220 + m87 & 0.0042 & @xmath221&@xmath141&@xmath222&@xmath223 + m84 & 0.0035 & @xmath224&@xmath225&@xmath226&@xmath227 + [ table ] | a new approach to constraining or determining the spin of a massive black hole is proposed .
a key parameter in the analysis is the dimensionless ratio , @xmath0 , of the energy released to the mass of the hole .
it is shown that the black hole spin @xmath1 may be written as a function of the dimensionless ratio of the black hole spin energy to the black hole mass .
when extraction of the black hole spin energy powers an outflow from the hole , the ratio @xmath0 provides an indication of the black hole spin . the method is applied to a sample of 19 very powerful classical double radio galaxies with a range of size and redshift .
the sources are found to have remarkably similar values of @xmath0 and @xmath1 , implying that the sources have very similar physical conditions at the time the outflow is generated .
the weighted mean value of the spins is @xmath2 . a sample of 29 central dominant galaxies is also studied ; most of these sources have amorphous or fri radio structure .
the sources have a broad range of ratios @xmath0 , and correspondingly broad range of black hole spins @xmath1 with values from about 0.001 to 0.4 , and a median value of about 0.03 .
the broad range probably results from the fact that each source is observed at a different stage in its lifetime .
the intrinsic range of parameters could be as tight as it is for the powerful classical double radio galaxies . |
motivated by the knill - laflamme - milburn scheme for scalable quantum computing with linear optics @xcite , in recent years the field of quantum interference and entanglement with photonic qubits has grown rapidly @xcite .
small numbers of entangled photons have been generated with stimulated parametric down - conversion @xcite and used for implementing various quantum information protocols @xcite .
more recently , several considerable methods for revealing macroscopic entangled states are reported with linear optics , such as by combining a single photon and a coherent beam on a beam splitter @xcite , amplifying and deamplifying a two - mode single photon entangled state @xcite , and so on . for quantum entanglement of a large number of photons @xcite , however , whether its fundamental principle or experimental demonstration is still a difficult and subtle task .
a cross - kerr nonlinear medium is capable of inducing an interaction between the photons @xcite , although its strength is very small for all experiments reported to date @xcite .
based on these weak nonlinearities one can implement photon - number quantum nondemolition measurement @xcite , entanglement detection @xcite , quantum logic gates @xcite , and miltiphoton entanglement @xcite . since the nonlinearities are extremely weak , it seems natural to improve experimental methods so as to produce large enough nonlinear phase shifts and then follow the previous schemes without bound in the limit of weak nonlinearities .
on the other hand , with current technology , it is also important to explore quantum circuit in the regime of weak nonlinearities for quantum information processing @xcite . in this letter
, we focus on the exploration of photon - number entangled states using weak nonlinearities . for each photon number @xmath0 , we show a quantum circuit to evolve two - mode signal photons , ranging from microscopic to macroscopic systems ( i.e. , from @xmath1 to @xmath2 ) . in the regime of weak nonlinearities ,
more importantly , we consider the maximal phase shift induced in the process of interaction between photons satisfying @xmath3 . moreover , in the absence of decoherence , we analyze error probability caused by the final homodyne measurement .
in fock space , consider an arbitrary two - mode @xmath0-photon - number state @xmath4 and @xmath5 , and @xmath6 and @xmath7 are complex parameters satisfying the normalization condition @xmath8 .
obviously , @xmath0-photon number state ( 1 ) includes some canonical entangled number states @xcite , such as single - photon entangled state and the noon state .
theoretically , state ( 1 ) can be conditionally produced by letting @xmath0 photons pass through a beam splitter and the two spatial modes @xmath9 , @xmath5 correspond to two outputs of beam splitter . throughout the subsequent context ,
let @xmath10-m$ ] with @xmath11 $ ] and we replace @xmath12 by @xmath13 for simplicity .
for a given photon number @xmath0 , we next show a method to detect the states @xmath13 by using weak nonlinearities .
suppose there exist @xmath0 photons traveling through two spatial modes @xmath9 and @xmath5 , namely signal modes ; and we introduce a coherent state @xmath14 in probe mode ; let @xmath15 and @xmath16 be respective phase shifts on the coherent probe beam according to the two signal modes , as shown in fig.1 . in order to avoid inducing @xmath17 in the interacting process
we herein introduce a single phase gate , i.e. , @xmath18 . after an @xmath19 homodyne measurement on the probe beam plus appropriate local phase shift operation @xmath20 on one of the signal modes using classical feed - forward information , at last
, the original state can be projected into one of the photon - number entangled states @xmath13 .
photons traveling through two spatial modes @xmath9 and @xmath5 ( say signal modes ) .
@xmath21 is a coherent state in probe mode . @xmath15 and @xmath16 are respective phase shifts on the coherent probe beam with several weak cross - kerr nonlinearities .
@xmath22 is a single phase gate used to evolve coherent state .
@xmath20 represents a phase shift on one of the signal modes based on the classical feed - forward information.,title="fig:",width=336 ] + we now describe our method in details . for @xmath0
is even , after the interaction between the photons and the action of the phase gate , the combined system @xmath23 evolves as @xmath24.\ ] ] after the @xmath19 homodyne measurement on the probe beam @xcite , the signal photons become @xmath25 } \}(a_{m}\texttt{e}^{-\texttt{i}\phi _ { m}(x ) } |n/2+m , n/2-m\rangle_{s_1s_2 } + b_{m}\texttt{e}^{\texttt{i}\phi _ { m}(x ) } |n/2-m , n/2+m\rangle_{s_1s_2}),\ ] ] where @xmath26 and @xmath27 \ { { x - 2\alpha \cos [ m(n-1 ) \theta ] } \ } \bmod 2\pi$ ] , @xmath28 .
the functions @xmath29 } \}$ ] , @xmath30 , are respective gaussian curves with peaks located at @xmath31 $ ] and these curves correspond to the probability amplitudes associated with the outputs of the signal photons .
@xmath20 are respective phase shift operations corresponding to the values of the @xmath19 homodyne measurement .
the midpoints between two neighboring peaks are designated as @xmath32 + \cos [ ( n/2-k)(n-1)\theta ] } \}$ ] , @xmath33 .
note that with these @xmath34 midpoint values @xmath35 , one can separate results of the homodyne measurement into @xmath36 intervals and then the input state can be projected into one of the states @xmath13 up to a phase shift operation @xmath37 on one of the signal modes .
clearly , for @xmath38 we observe immediately the output state @xmath39 ; for @xmath40 , we obtain the states @xmath41 ; and for @xmath42 we obtain the state @xmath43 .
considering there exist small overlaps between two neighboring curves , the error probabilities are thus given by @xmath44 , where @xmath45 - \cos [ ( n/2-k)(n-1)\theta ] } \ } \approx ( n-2k-1)(n-1)^2\alpha { \theta ^2}$ ] are the distances of two nearby peaks . similarly , for @xmath0 is odd , the combined system then evolves as @xmath46.\ ] ] after the measurement on the probe beam , the signal photons become @xmath47\ } \nonumber \\ & & \times[a_{m}\texttt{e}^{-\texttt{i}\phi_{(2m+1)/2}(x)}|(n+1)/2+m,(n-1)/2-m\rangle_{s_1s_2 } + b_{m}\texttt{e}^{\texttt{i}\phi _ { ( 2m+1)/2}(x)}|(n-1)/2-m,(n+1)/2+m\rangle_{s_1s_2}],\end{aligned}\ ] ] where in order to simplify the notations we make use of the same expressions as the case of even @xmath0 for the functions @xmath48\}$ ] and @xmath49 .
the peaks of gaussian curves locate at @xmath50 $ ] .
then , we can derive the same results of the midpoints @xmath35 between two neighboring peaks and the distances @xmath51 separated by two nearby peaks .
by now , in the regime of weak kerr nonlinearities we have described the evolution of @xmath0 photon combined system .
we next discuss some intriguing applications of the exploration of photon - number entangled states .
in many quantum measurements for linear optical quantum computation , one should always attempt to detect the signal photons and then project them into a desired subspace . surprisingly , for a given photon number @xmath0 , the setup in our scheme can be used for these processes based on postselection and classical feed - forward ( an additional manipulation ) .
so , a direct application of the present scheme leads to an entangling gate for the states @xmath52\}$ ] , and especially , for @xmath53 it yields one of the maximally entangled number states @xmath54 .
moreover , it is easy to see that the present scheme is suitable for constructing two - qubit polarization parity gate ( see fig.2 in ref.@xcite ) , discriminating between @xmath55 and @xmath56 , nondestructively ( see fig.1 in ref.@xcite , also see fig.2 in ref.@xcite ) , and so on .
another important application of the present scheme is as an analyzer for the states @xmath57 . consider a state belonging to the set of states @xmath57 in signal modes .
after the evolution of a series of optical devices followed by an @xmath19 homodyne measurement on the probe beam , based on the value of the measurement one can infer immediately what the input must have been with a small error probability .
then , a conditional phase shift operation on one of the modes is necessary to restore the output state to that identified .
in other words , the suggested analyzer of the states @xmath57 is nondestructive and thus the unconsumed signal photons can be recycled for further use .
note that the cross - kerr nonlinearities are extremely weak and the order of magnitude of them is only @xmath58 even by using electromagnetically induced transparency @xcite . in the present scheme ,
let @xmath59 and then obtain @xmath60}/2 $ ] .
therefore , by applying an appropriate coherent probe beam the present scheme can has a small enough error probability and then be realized in a nearly deterministic manner . given @xmath61 with @xmath62 , for example , then we have @xmath63 .
clearly , let @xmath1 and mean photon numbers of coherent state @xmath64 , then the above value of error probability holds . also , when @xmath65 and letting @xmath66 we can also have the given error probability . in summary , we show an architecture of exploration of photon - number entangled states using weak nonlinearities . also , we suggest some interesting applications of the present scheme and analyze its error probabilities .
the present scheme has two remarkable advantages .
first , our scheme is feasible with the current experimental technology , because there is no large phase shift ( @xmath17 with @xmath67 , for example ) in the interacting process with weak kerr nonlinearities and then the strength of the nonlinearities we required are orders of magnitude in current practice .
second , by analyzing the error probability we show that our scheme works in a nearly deterministic way .
this work was supported by the national natural science foundation of china under grant nos : 11475054 , 11371005 , hebei natural science foundation of china under grant no : a2014205060 , the fundamental research funds for the central universities of ministry of education of china under grant no:3142014068 , langfang key technology research and development program of china under grant no : 2014011002 .
e. knill , r. laflamme , and g. j. milburn , nature ( london ) * 409 * , 46 ( 2001 )
. j. w. pan , z. b. chen , c. y. lu , h. weinfurter , a. zeilinger , and m. @xmath68kowski , rev .
mod . phys . * 84 * , 777 ( 2012 ) .
p. g. kwiat , k. mattle , h. weinfurter , a. zeilinger , a. v. sergienko , and y. shih , phys .
lett . * 75 * , 4337 ( 1995 ) . c. simon and d. bouwmeester , phys .
lett . * 91 * , 053601 ( 2003 ) .
p. kok , w. j. munro , k. nemoto , t. c. ralph , j. p. dowling , and g. j. milburn , rev .
* 79 * , 135 ( 2007 ) .
p. sekatski , n. sangouard , m. stobi@xmath69ska , f. bussi@xmath70res , m. afzelius , and n. gisin , phys .
a * 86 * , 060301 ( 2012 ) .
r. ghobadi , a. lvovsky , and c. simon , phys .
. lett . * 110 * , 170406 ( 2013 ) .
a. lamas - linares , j. c. howell , and d. bouwmeester , nature ( london ) * 412 * , 887 ( 2001 ) .
h. s. eisenberg , g. khoury , g. a. durkin , c. simon , and d. bouwmeester , phys .
* 93 * , 193901 ( 2004 ) .
f. de martini , f. sciarrino , and c. vitelli , phys .
lett . * 100 * , 253601 ( 2008 ) .
p. sekatski , n. gisin , and n. sangouard , phys .
lett . * 113 * , 090403 ( 2014 ) .
n. imoto , h. a. haus , and y. yamamoto , phys .
a * 32 * , 2287 ( 1985 ) .
w. j. munro , k. nemoto , and t. p. spiller , new j. phys .
* 7 * , 137 ( 2005 ) .
t. guerreiro , a. martin , b. sanguinetti , j. s. pelc , c. langrock , m. m. fejer , n. gisin , h. zbinden , n. sangouard , and r. t. thew , phys .
lett . * 113 * , 173601 ( 2014 ) .
h. schmidt and a. imamo@xmath71lu , opt
* 21 * , 1936 ( 1996 ) .
m. d. lukin and a. imamo@xmath71lu , phys .
. lett . * 84 * , 1419 ( 2000 ) .
n. matsuda , r. shimizu , y. mitsumori , h. kosaka , and k. edamatsu , nature photonics * 3 * , 95 ( 2009 ) .
w. j. munro , k. nemoto , r. g. beausoleil , and t. p. spiller , phys .
a * 71 * , 033819 ( 2005 ) .
s. d. barrett , p. kok , k. nemoto , r. g. beausoleil , w. j. munro , and t. p. spiller , phys .
rev . a * 71 * , 060302 ( 2005 ) .
y. b. sheng , f. g. deng , and g. l. long , phys .
a * 82 * , 032318 ( 2010 ) .
d. ding and f. l. yan , acta phys . sin . *
62 * , 100304 ( 2013 ) .
k. nemoto and w. j. munro , phys .
* 93 * , 250502 ( 2004 ) .
p. kok , phys .
a * 77 * , 013808 ( 2008 ) .
d. ding , f. l. yan , and t. gao , j. opt .
b * 30 * , 3075 ( 2013 ) .
b. t. kirby and j. d. franson , phys .
a * 87 * , 053822 ( 2013 ) .
t. wang , h. w. lau , h. kaviani , r. ghobadi , and c. simon , arxiv : quant - ph/1412.3090 . y. q. he , d. ding , f. l. yan , and t. gao , j. phys .
b : at . mol .
48 * , 055501 ( 2015 ) .
q. lin , b. he , j. a. bergou , and y. h. ren , phys .
a * 80 * , 042311 ( 2009 ) .
d. ding and f. l. yan , phys .
a * 377 * , 1088 ( 2013 ) .
d. ding , f. l. yan , and t. gao , sci .
china - phys .
astron . * 57 * , 2098 ( 2014 ) .
f. l. yan , t. gao , and e. chitambar , phys .
a * 83 * , 022319 ( 2011 ) .
t. gao , f. l. yan , and s. j. van enk , phys .
lett . * 112 * , 180501 ( 2014 ) .
a. n. boto , p. kok , d. s. abrams , s. l. braunstein , c. p. williams , and j. p. dowling , phys . rev
. lett . * 85 * , 2733 ( 2000 ) . | a method for exploring photon - number entangled states with weak nonlinearities is described .
we show that it is possible to create and detect such entanglement at various scales , ranging from microscopic to macroscopic systems . in the present architecture , we suggest that the maximal phase shift induced in the process of interaction between photons is proportional to photon numbers . also , in the absence of decoherence we analyze maximum error probability and show its feasibility with current technology . |
the standard poisson - boltzmann equation ( pb ) is accurate for dilute aqueous solutions containing small monovalent ions .
the first approximation of the pb theory is the mean - field account of the electrostatic interactions , where the discrete nature of ions is neglected and each ion interacts with the mean - potential generated by the local charge density distribution @xmath0 according to the coulomb law , @xmath1 , where @xmath2 is the dielectric constant .
this description clearly ignores ionic correlations .
the other approximation within the pb theory is the lack of internal structure of each ion : an ion is a point in space characterized only by its charge . under this reduced description
the ions f@xmath3 , cl@xmath3 , and no@xmath4 are indistinguishable , although their size @xcite and polarizability @xcite are obviously different .
the pb description , therefore , breaks down when the above reductions can no longer be justified . in the present work we consider the weakly correlated limit so that the mean - field description of electrostatics holds , but we go beyond the standard pb equation by including the nonelectrostatic excluded volume effects @xcite . in aqueous solutions
the excluded volume interactions are further increased on account of hydration , the binding of water molecules to ions , so that the effective ionic radii are larger than the crystallographic ones @xcite .
another effect of hydration , related directly to electrostatics , is the dielectric decrement @xcite associated with the decrease of the medium dielectric constant on account of polarization saturation of a solvent constituting the hydration shell . in the present work we do not consider this effect and focus strictly on the non - electrostatic excluded volume effects .
the excluded volume effects reorganize the double layer so that a density profile is no longer monotonically decreasing , but it acquires oscillatory structure reflecting molecular composition of an electrolyte @xcite . clearly , the modifications of the density profile lead to modified electrostatic properties . in the present study
we take a close look into the excluded volume effects where ions are represented as hard spheres with the same diameter , the so - called , restricted primitive model . not only
do we investigate how the ionic system responds to these effects under various conditions , but also how well various theories reproduce this behavior .
the hard - sphere interactions lead to the nonlocal effects , demanding nonlocal functionals to account for the oscillatory structure in the density profile .
the nonlocal treatments , such as the density functional theory ( dft ) , can be computationally demanding if the system is sufficiently complex . frequently in electrostatics ,
a variant of the local density approximation ( lda ) is used , where the ideal gas entropy is substituted by that of the lattice - gas leading to the modified poisson - boltzmann equation ( mpb ) @xcite . with growing popularity of the mpb as a simple and tractable model
, we feel that a careful and systematic study of this method is still lacking .
first of all , the mpb is a local theory inappropriate for moderately and highly inhomogeneous fluids .
furthermore , even as a theory of a homogeneous fluid it already fails at predicting the correct second virial term .
we consider another variant of the lda based on the scaled particle theory ( spt ) .
surprisingly , although highly accurate for treating homogeneous fluids , it is less accurate than the mpb , which is unreliable for homogeneous fluids .
we compare all the results against the nonlocal fundamental measure density functional theory @xcite .
our main focus is the region near and at the contact with the wall , where exact thermodynamic relations are known .
we investigate the scaling properties of the contact quantities and propose an analytical fit to the dft data that permits an accurate prediction of these quantities , without resorting to numerical computations involving the entire region . throughout the paper
we make frequent use of the initials : dft , lda , mpb , and spt . to avoid confusion
we give a quick account of each .
dft always refers to the fundamental measure density functional theory as originally derived by rosenfeld @xcite .
this is the only nonlocal theory used in this work .
lda refers to the general class of approximations where a density functional is defined locally based on a free energy of a homogeneous fluid .
mpb is the special example of the lda , when the excluded volume effects are represented as that of a homogeneous lattice gas @xcite .
and the spt stands for the scaled particle theory and refers to the theoretical approach based on scaling a particle size @xcite .
we use spt to refer to another lda approximation based on the spt results for a uniform fluid . in the paper two lda approximations
will be considered : the mpb and the spt . the paper is organized as follows . in section
ii we discuss a reversible work of inserting a test particle into a hard sphere fluid , @xmath5 , and emphasize its nonlocal character by deriving the exact result for inserting a point particle into a hard - sphere fluid .
we then introduce the fundamental density functional theory as an accurate nonlocal treatment of the hard - sphere interactions . in section iii
we review the contact value theorem and explore its implications for the counterion distribution near a hard charged surface .
we point out that the lda theories lead to contact relations different from the exact contact value theorem .
this is the result of representing the overcrowding in the double - layer as the density `` saturation '' . in section
iv we investigate the scaling of the quantity @xmath6 , which represents the reversible work of brining the uncharged fluid particle from the bulk into contact with the wall .
we propose the analytical formula for @xmath6 that fits the dft data points . in section
v we apply this formula to calculate the surface potential and the differential capacitance . section vi finalizes this work with the concluding remarks .
we write the free energy functional as having three separate contributions , @xmath7 the ideal gas contribution , @xmath8 the mean - field electrostatic interactions , @xmath9 [ \rho_c({\bf r}')+\rho_{\rm f}({\bf r}')]}{|{\bf r}'-{\bf r}|},\ ] ] and the excluded volume contributions , @xmath10 , where all ions are taken to have the same diameter @xmath11
. @xmath12 denotes a fixed charge of a macromolecule and in the present work we consider a charged wall , @xmath13 , where @xmath14 is the surface charge .
@xmath15 is the charge density of mobile ions , where @xmath16 is the charge of an ion species @xmath17 and @xmath18 is the number of all ionic species .
finally , @xmath19 is the de broglie wavelength .
the equilibrium density minimizes the grand potential functional , @xmath20 with respect to a density distribution , @xmath21 , and leads to @xmath22}. \label{eq : rho_j}\ ] ] where @xmath23 is the mean electrostatic potential , @xmath24 , and @xmath25 is the bulk density of an ion species @xmath17 .
physically , @xmath5 represents the reversible work of inserting a test particle at position @xmath26 into a hard - sphere fluid @xcite .
if the system has a wall , the insertion performed far from the wall , @xmath27 is identical with the excess ( over the ideal gas ) chemical potential .
alternatively , @xmath5 can be seen as a depletion interaction : as a particle approaches a hard - wall , it feels itself being pushed towards the wall by the other particles .
application of the poisson equation , @xmath28 , leads to a kind of modified poisson - boltzmann equation , @xmath29},\ ] ] where the surface charge @xmath14 is determined by the boundary conditions at a charged wall , @xmath30 in order to obtain the potential , one still needs some kind of closure for @xmath5 that will determine the excluded volume effects .
@xmath5 is a nonlocal quantity . to see the origins of the non - locality we consider a simple case where the insertion work can be expressed as an exact density functional
the case we refer to is the reversible work of inserting a _ point test particle _ into a hard - sphere fluid , @xmath31 .
the range of the hard - core interaction between the test point particle and the fluid hard - spheres is half the diameter , @xmath32 .
the reversible work is related to the insertion probability , @xmath33 . for a given instantaneous configuration of a hard sphere fluid ,
the space is either occupied or is empty .
the point particle can be inserted into the empty space between the hard spheres .
the instantaneous insertion probability , @xmath34 , is either zero , if we hit on an occupied space , or one , if a cavity is found , @xmath35 where @xmath36 is the density operator , @xmath37 is the total number of fluid particles , and @xmath38 is the heaviside step function .
averaging over all the configurations , we obtain @xmath39 the reversible work of inserting a point particle at position @xmath26 is @xmath40 .
\label{eq : w_sigma/2}\ ] ] the non - locality of @xmath31 is already evident in this simple example and shows that the work of inserting a point particle depends on the weighted average of the inhomogeneous density distribution .
the extent of the nonlocality is determined by the range of the hard - core interaction . to obtain an expression for @xmath5 ,
a reversible work associated with expanding a point particle to diameter @xmath11 has to be considered , and @xmath31 constitutes only a part of the entire quantity @xmath5 . for the homogeneous fluid , this procedure leads to the scaled particle theory ( spt ) @xcite , in which the work of expansion to finite diameter @xmath11 is interpolated between the two exact limits , @xmath41 and @xmath42 , where @xmath43 is the range of the hard sphere interaction between a test and a fluid particle . for an inhomogeneous fluid
the derivation is more involved @xcite .
the ideas of the spt applied to inhomogeneous fluids lead to the formulation of the fundamental measure density functional theory ( dft ) @xcite .
the dft approximates the excess free energy due to the hard - sphere interactions through the free energy contribution , @xmath44=\int d{\bf r}\,\phi_{\rm hs}$ ] , where @xmath45 the hard sphere contributions are locally represented through weighted densities , instead of a physical density , as is the case for the lda approximation , @xmath46 with @xmath47 denoting a weight function @xcite .
there are five weight functions which are obtained through decomposition of the mayer-@xmath48 function into convolution products @xcite .
the two of the weight functions are vector quantities and represent a distribution of normal to the sphere surface vectors . in the uniform fluid they have no contribution .
the weight functions are given below , @xmath49 and @xmath50 , @xmath51 , and @xmath52 , where @xmath53 .
since we assume the size of all ions to be the same , the weight functions are the same for all the ionic species .
the insertion work @xmath5 is defined as the functional derivative of the excess part of free energy , @xmath54 the dft theory presented above accurately captures the dimensional crossover when the 3d fluid is confined to a quasi-2d space @xcite if a degree of freedom is completely eliminated in the transverse direction , a 2d fluid is accurately recovered .
although highly accurate , see fig .
( [ fig : rho_simulation ] ) , the numerical implementation of the fundamental measure density functional theory is quite involved . to simplify the calculations , local density approximations ( lda )
have often been invoked . within the lda ,
the thermodynamic relations derived for a homogeneous fluid are applied _ locally _ to an inhomogeneous situation .
for example , within the lda the insertion probability @xmath5 is supposed to be a function of the local density @xmath55 and is identified with the `` local '' excess chemical potential , @xmath56 . clearly
the specific form of the lda will depend on the approximate form of @xmath5 .
the mpb @xcite relies on @xmath5 derived from the lattice - gas description of electrolyte , @xmath57 , where @xmath58 is the packing fraction .
the spt expression is @xmath59 .
in this work we focus on the region at a contact with a charged wall . in this region the excluded volume effects are most pronounced since the ion concentration is highest , and for many problems the contact quantities are of primary interest .
this is the case for calculation of the differential capacitance @xcite , or the estimation of the strength of the electrostatic interactions between colloids trapped at an interface @xcite .
the theoretical investigation in this region is facilitated by the existence of exact thermodynamic relations .
this section reviews some analytical results valid at a contact with a wall . throughout the paper
we make frequent use of either the subscript or the superscript @xmath60 to indicate that a quantity is taken at contact with the wall . in the same way we use @xmath61 to
indicates bulk values .
we remind that the charge and number density are @xmath15 and @xmath62 , respectively . consequently , the bulk and contact value of the number density is @xmath63 and @xmath64 , respectively .
the simplest expression of the contact value theorem relates the bulk pressure @xmath65 to the average momentum transferred at the hard - wall : @xmath66 , where @xmath67 is the fluid density at the contact with the wall .
if the hard - wall has a surface charge @xmath14 and the particles are hard - spheres with a central charge , then the contact value theorem takes the form @xcite , @xmath68 within the dft formalism @xmath69 the pressure is given according to the scaled particle theory .
this result is a highly accurate approximation and the contact value theorem according to the dft is essentially exact .
the contact value relation is the result of the mechanical force balance condition , @xmath70 = \rho_c{\mbox{\boldmath $ \nabla$}}\psi + \rho{\mbox{\boldmath $ \nabla$}}w_{\rm hs } , \label{eq : force_bal}\ ] ] where the thermodynamic force on the left - hand side is counterbalanced by the electrostatic force and the depletion force due to the collisions with other fluid particles on the right - hand side .
the explicit expression for the contact density , using eq .
( [ eq : rho_j ] ) , is @xmath71e^{-\delta_w } \label{eq : rho_w}\ ] ] where we have introduced the quantity @xmath72 representing the reversible work expanded to bringing a hard - sphere particle ( with no electric charge ) from the bulk electrolyte to the wall .
this is not to say that @xmath6 is independent of electrostatic properties .
the work of bringing an uncharged hard - sphere to the wall depends on the ionic distribution within the double - layer that depends on the electrostatic interactions . substituting the density in eq .
( [ eq : rho_w ] ) into the contact value theorem gives @xmath73e^{-\delta_w } = p + \frac{\sigma_c^2}{2\epsilon } , \label{eq : cvt11}\ ] ] for the case @xmath74 , @xmath75 is known exactly and is determined by the bulk quantities and the surface charge .
the standard pb equation describes this scenario .
a finite @xmath6 modifies the contact potential .
@xmath76 signifies that it is easier to insert a particle far away from the wall than at the contact .
this is intuitive because near the wall counterions are in excess .
consequently the contact potential is larger than that obtained from the standard pb equation steric interactions deplete some ions from the first layer making it more difficult to neutralize the wall charge . on the other hand , @xmath77 signifies that it is easier to insert a particle at the wall than far from it .
although this seems counterintuitive , it can occur at low values of the surface charge @xmath14 and/or large @xmath78 . to see this we consider the limit @xmath79 , @xmath80 in this limit
, @xmath6 is always negative ( since @xmath81 ) , and the larger the bulk packing fraction @xmath78 , the larger will be the modulus of @xmath6 . at large ionic concentrations such as , for example ,
found in ionic liquids the range of @xmath14 where @xmath6 is negative can be significant . in fig .
( [ fig : delta_vs_eta ] ) we plot @xmath6 as a function of @xmath78 for different values of the surface charge .
the crossover where @xmath6 changes sign to negative is pushed to larger @xmath78 as @xmath14 becomes larger .
eventually the crossover is suppressed as @xmath78 passes the freezing transition ( at @xmath82 ) . within the local density approximation
the insertion work is a local quantity , @xmath83 and the force density due to the hard - sphere interactions can be related to the gradient of a local pressure , @xmath84=-{\mbox{\boldmath $ \nabla$}}[p - k_bt\rho_b].\ ] ] the force balance equation in eq .
( [ eq : force_bal ] ) becomes @xmath85 leading , after integrating over a half - space , to the contact relation , @xmath86 where @xmath87 is the local pressure at contact with the wall . the notion of a local pressure
can be somewhat misleading , since the pressure is normally associated with bulk properties .
it is only sensible to talk about a local pressure within the framework of a local density approximation .
going back to the exact contact value theorem in eq .
( [ eq : cvt ] ) shows that the correct thermodynamic result is obtained only if @xmath88 .
paradoxically , if one tries to improve on this result and go beyond the pb description by including an excess free energy due to electrostatic or steric correlations using a lda formalism , one obtains less accurate contact value density than if the correlations are neglected altogether .
the unphysical contact value relation for the lda treatment is a consequence of an unphysical density profile .
the lda approximation fails to capture the oscillating discrete structure of a fluid , see fig .
( [ fig : rho_dft ] ) .
as seen in that figure , the lda approximations account for overcrowding through `` density saturation '' , while a realistic description shows overcrowding as layering of a fluid near a wall .
it is tempting to assume that the `` saturation '' observed in a density for the lda models represents a weighted rather than a physical density profile , @xmath89 where @xmath90 is some weight function , so that oscillations are completely smoothed out .
such conjecture , however , entails that @xmath91 obtained from the lda is also not physical but only a weighted quantity .
we , therefore , consider that all quantities produced by the lda are what they are , and that the saturation of a density profile is an artifact of the lda treatment .
to test accuracy of the lda one should rather consider other quantities such as @xmath91 , which is of direct interest to most problems of electrostatics .
below we look into details of two lda models . as mentioned earlier , the spt results are highly accurate for homogeneous fluids , producing the exact second virial coefficient .
furthermore , the third and higher order virial coefficients are very accurate , @xmath92.\ ] ] on the other hand , the lattice - gas model ( from which the mpb is derived ) already fails at the second virial coefficient , @xmath93,\ ] ] and is unfit for treating homogeneous fluids . on the other hand , when the lda is applied to inhomogeneous fluids ,
it is hard to predict what will happen . in fig .
( [ fig : phi_vs_sigma_c_lda ] ) we plot the contact potential for various models .
first we note that the difference between the pb and the dft results . the finite size effects lead to an increase of the surface potential , and the larger the surface charge , the larger the discrepancy , since more counterions are crowded into the double - layer .
next we discuss the lda models .
the mpb model underestimates while the lda based on the spt overestimates the dft result .
we next consider the quantity @xmath6 as defined in eq .
( [ eq : delta_w ] ) . as shown in fig .
( [ fig : delta_lda3 ] ) the mpb reproduces this quantity better than the stp .
the advantage of the mpb model is its analytical tractability , and according to this model @xmath6 yields a very simple expression , @xmath94 with @xmath6 parabolic in @xmath14 .
all the calculations in this work are carried out for the @xmath95 symmetric salt , so that the bulk number density is @xmath96 , where @xmath97 denotes the bulk salt concentration . in this part of the study
we investigate the scaling of the quantity @xmath6 .
@xmath6 is a function of three dimensionless variables : @xmath98 , @xmath99 , and @xmath100 , corresponding to volume , surface , and length , respectively .
@xmath101 represents the packing fraction of counterions projected on a 2d plane .
the 2d solid phase transition is at @xmath102 , and the maximum possible packing fraction is at @xmath103 .
if @xmath104 , potentially all the counterions can collapse onto the charged surface , if attraction is sufficiently strong ( although @xmath105 would involve a phase transition as well ) .
if , on the other hand , @xmath106 , collapse onto a single plane is prevented and a second fluid layer must form . in the reduced units , the contact value theorem is @xmath107 where @xmath108 is the reduced potential , and @xmath109 is the volume of a single fluid particle . according to the lda
, @xmath6 is a function of two independent parameters , @xmath78 and @xmath110 .
later , as we investigate @xmath6 for various parameters using the dft , we do not find this reduced description to be confirmed , and we conclude that it is an artifact of the lda . furthermore , this implies that the scaling of the contact potential @xmath75 according to the lda description is described by two independent parameters .
the two parameter scaling of the lda can be ascertained from the contact relation in eq .
( [ eq : cvt_lda ] ) . to simplify things we first consider the dilute limit , so as to suppress the contributions from the bulk eliminating the dependence on @xmath78 , @xmath111 .
the main results are in fig .
( [ fig : delta_vs_sigma_c_a ] ) and fig .
( [ fig : delta_vs_lambda ] ) .
the figures show the scaling behavior of @xmath6 .
the general observation is that the scaling with the surface charge is stronger than that with the bjerrum length .
this can be understood since the surface charge controls both the interactions between the surface and counter charges and the counterion concentration . on the other hand ,
the bjerrum length regulates only the strength of the ionic interactions .
our aim is to propose a fit which reproduces the scaling of @xmath6 .
the desirable procedure would be to suggest an expression based on a physically motivated model .
unfortunately , an accurate physically motivated fit could only be found for a limited range . to accurately parametrize the entire range we had to resort to a fitting procedure where the choice of a functional form was based on simplicity and convenience .
the following functional form was found convenient , @xmath112\}^\alpha$ ] .
fully parametrized , this form reads @xmath113^{3/4}. \label{eq : delta0}\end{aligned}\ ] ] the accuracy of this fit is demonstrated in figs .
( [ fig : delta_vs_sigma_c_a ] ) and fig .
( [ fig : delta_vs_lambda ] ) in comparison to the dft data points .
we also compare this formula with the simple expression from the mpb model in eq .
( [ eq : delta_mpb ] ) which in reduced units reads @xmath114 there is a curious similarity between the square term in our fit @xmath115 and the coefficient @xmath116 in the mpb expression .
if this similarity is more than coincidental , it would suggest that although reduced , the mpb description accurately captures some aspects of the hard - sphere interactions .
we now turn to investigate the contributions of finite concentrations .
we suggest the following formula , @xmath117 so that @xmath6 as a function of @xmath101 is merely shifted by the quantity @xmath118 .
this fit is suggested by fig .
( [ fig : delta_vs_sigma_c_b ] ) where we compare @xmath6 for finite @xmath78 with that for a dilute limit . in the relevant regime @xmath6 as a function of @xmath101 is offset by a value independent of @xmath101 .
we find the following accurate fit to the function @xmath118 , @xmath119 .
\label{eq:}\ ] ] we subtract @xmath120 from @xmath6 so that the quantity @xmath121 is always positive , see eq .
( [ eq : delta_dilute ] ) and fig .
( [ fig : delta_vs_eta ] ) . to ensure that @xmath122 , so that @xmath123 for all values of @xmath78 , we apply the form , @xmath124
we now have the complete expression for @xmath6 , @xmath125 where @xmath121 is given in eq .
( [ eq : delta_w_complete ] ) . to obtain the contact potential , we insert @xmath6 into eq .
( [ eq : cvt11 ] ) to get @xmath126 in fig .
( [ fig : phi_vs_sigma_c_a ] ) and fig .
( [ fig : phi_vs_sigma_c_b ] ) we compare the exact dft results for the contact potential with the ones obtained using the analytical scaling function eq .
( [ eq : coshphi ] ) and with the predictions of the mpb model .
the scaling form derived in the present paper agrees very well with the contact potential obtained using the exact dft , even at densities all the way to the solid phase transition . for largest particle sizes , fig .
( [ fig : phi_vs_sigma_c_b ] ) shows that the scaling function slightly disagrees with the dft results .
this happens when @xmath127 and a third layer of condensed counterions starts to form .
a quantity that is of particular interest for electrochemistry is the differential capacitance of the double - layer @xcite , @xmath128 the standard pb result gives @xmath129 which is the , so - called , gouy - chapman capacitance .
@xmath130 is the inverse debye length . at low @xmath131
this formula works well , but as @xmath131 increases , steric repulsion lowers the capacitance as counterions are prevented from accumulating near the electrode , leading to a `` camel - shaped '' curve of fig .
( [ fig : capacitance ] ) .
the scaling function derived in the present paper allows us to easily calculate the differential capacitances , which once again are in a very good agreement with the dft .
we have investigated the effects of steric interactions on the ionic distribution and electrostatic properties near a charged hard wall .
we have specifically focused on the reversible work required to bring a hard spherical particle ( with no electric charge ) from the bulk to the wall , @xmath6 .
the quantity @xmath6 is nonlocal and it controls the steric effects in the double - layer . typically @xmath76 .
this implies that the overcrowding near a wall causes the depletion of counterions away from the wall , leading to the increased contact potential .
the less common case , @xmath77 , represents the situation where counterions are forced from the dense bulk toward the wall .
this effect decreases the value of the contact potential .
we have also studied the scaling behavior of @xmath6 calculated using the nonlocal dft theory .
we expected that the careful look into the scaling behavior can reveal a simple underlying mechanism .
unfortunately , no simple model could cover the entire domain .
instead we proposed a functional form to fit the dft data points for @xmath6 .
such a fit allows for simple and accurate calculation of electrostatic properties at contact with the wall without resorting to numerical schemes of an entire region .
finally , we have explored the discrepancies between the lda and exact description .
the lda description shows that @xmath6 depends on two free parameters , in comparison to the three parameters given by full physical description ( dft ) .
another artifact of the lda is the `` saturation '' seen in a density profile rather than oscillations due to layering .
this saturation effect leads to unphysical contact value relation .
for various lda models , we find that the mpb based on the lattice - gas model is more accurate than the lda based on the spt theory , which for treating a uniform fluid is extremely accurate .
this indicates that some cancellation of errors is at work .
we remind that our analysis is limited to the weak - coupling limit , where electrostatic interactions are accurately captured by the mean - field treatment . to accurately account for the correlations together with the hard - sphere interactions
remains a great challenge .
it is our hope to treat the strong - coupling limit of the hard - sphere ions in the future .
this work was partially supported by the cnpq , fapergs , inct - fcx , and by the us - afosr under the grant fa9550 - 09 - 1 - 0283 .
we thank alexandre pereira dos santos for furnishing us with the simulation data in fig .
( [ fig : rho_simulation ] ) . | we explore the effect of steric interaction on the ionic density distribution near a charged hard wall . for weakly charged walls , small particles , and monovalent ions
the mean - field poisson - boltzmann equation provides an excellent description of the density profiles . for large ions and large surface charges ,
however , deviations appear . to explore these
, we use the density functional theory .
we find that local density functionals are not able to account for steric interactions near a wall .
based on the weighted density approximation we derive a simple analytical expression for the contact electrostatic potential which allows us to analytically calculate the differential capacitance of the double layer . |
as it is known , despite of the fact that magnetism is an essentially quantum effect , a wide range of magnetic phenomena can be successfully described in the framework of classical models .
one of the most widely used of such models is the one by landau and lifshitz , when magnetic is considered in the continuous limit and interaction between magnetic dipoles is taken into account in terms of some effective magnetic field .
the simplest case of the landau lifshitz model is the case of the so - called isotropic continuous heisenberg ferromagnetic spin system , which is governed by the equation @xmath2 , \qquad \vec s^{2 } = 1 .
\label{heis}\ ] ] here @xmath3 and @xmath4 is the two - dimensional laplacian , @xmath5 .
this equation attaches much attention not only from the viewpoint of its application in the physics of magnetic phenomena , but also from the viewpoint of the theory of integrable nonlinear partial differential equations .
it is known that in the @xmath6-dimensional case , @xmath7 , \label{heis11}\ ] ] this equation , which has been discussed in a large number of publications ( see @xcite as well as the books @xcite and references therein ) , can be solved using the inverse scattering transform ( ist ) .
another well studied reduction of ( [ heis ] ) is the static two - dimensional case , which may be referred to as @xmath8-dimensional one , @xmath9 = 0 , \label{heis02}\ ] ] ( see , e.g. , @xcite ) and which is closely related to the elliptic sine - gordon model .
the subject of the present paper are the stationary structures of the isotropic two - dimensional classical continuous heisenberg spin system and we look for solutions of ( [ heis ] ) which are of the form @xmath10 ( the so - called tijon
wright ansatz with zero frequency @xcite ) . of course , this reduction ( like any other ansatz ) is a necessity , if we want to proceed analytically , and is due to the fact that we can not at present integrate the original @xmath0-dimensional equation . on the other hand , the stationary structures which are discussed below are of much interest for the physics of magnetism and nonlinear physics in general since they are realization of the so - called dynamical solitons . in physics of magnetic phenomena there
exist two types of localized structures .
first is the domain walls ( kinks ) which connect two different ground states and which can not be destroyed without remagnetising regions of macroscopic sizes ( that is why they are called ` topological solitons ' ) .
another type is solitons , which can be viewed as bound states of a huge number of magnons .
the question of their existence and stability , contrary to the case of the topological solitons , is not so trivial . in some sense , they exist due to the presence of some conserved quantities such as number of magnons , total momentum and energy @xcite .
these are the only physical constants of motion which do not depend on the model we use ( as to an infinite number of conservation laws appearing , e.g. , in @xmath6-dimensional landau lifshitz equation , they seem to be an attribute of the model and do not survive when we move to a more realistic one ) .
the moving stationary structures are the simplest field configurations for which all of the constants are non - zero , i.e. they are the simplest of general ( or non - degenerate ) ones .
studying these structures one can explicitly see the interaction of the nonlinearity and dispersion which is known to be the core mechanism of creation of localized objects ( solitons ) not only in magnetic systems but in many other areas of the nonlinear physics .
after substitution ( [ stationary ] ) equation ( [ heis ] ) becomes @xmath11 + \left ( \vec v , \nabla \vec s \right ) = 0 .
\label{mainvec_eq}\ ] ] introducing the variables @xmath12 , \qquad \bar z = \frac{v}{4 g } \left [ x - v_{x}t - i ( y - v_{y}t ) \right]\ ] ] where @xmath13 , we can rewrite it as @xmath14 + \lambda^{2 } \vec s_{z } + \lambda^{-2 } \vec s_{\bar z } = 0 \label{tijon}\ ] ] or , equivalently , as @xmath15 = 0 \label{veczz}\ ] ] with @xmath16 , where the angle @xmath17 is defined by @xmath18 , @xmath19 .
equation ( [ tijon ] ) is known to be integrable ( its zero - curvature representation ( zcr ) one can find in the paper @xcite ) , and one can tackle it by elaborating the corresponding inverse scattering transform .
however , in the present paper we do not discuss this question .
our aim is to establish the relations between the model considered and the other integrable models , which will provide us with a wide range of physically interesting solutions . for our further purposes
it is convenient to rewrite ( [ veczz ] ) in the matrix form using the correspondence @xmath20 where @xmath21 are the pauli matrices @xmath22 equation ( [ veczz ] ) then can be presented as @xmath23 = 0 . \label{main_eq}\ ] ] namely this equation is the central object of our investigation .
we will use the following remarkable fact : equation ( [ main_eq ] ) is gauge equivalent to the @xmath24 nonlinear @xmath25-model @xcite in the similar way as , e.g. , the @xmath6-dimensional classical continuous heisenberg ferromagnetic spin system ( [ heis11 ] ) is equivalent to the nonlinear schrdinger equation ( see @xcite ) , or the ishimori magnetic @xcite to the davey stewartson system ( see @xcite ) .
this equivalence can be briefly described in terms of the ist as follows : some combinations of the jost functions of the linear problem associated with the @xmath24 @xmath25-model solve equation ( [ main_eq ] ) ( below we shall discuss this question more comprehensively ) .
the fact that model ( [ veczz ] ) , or ( [ main_eq ] ) , is related to the @xmath24 nonlinear @xmath25-model is a generalization of the already known result ( see , e.g. , @xcite ) that in the static case equation ( [ heis02 ] ) is gauge equivalent to the elliptic sine - gordon equation .
the @xmath24 @xmath25-model , as it has been shown in @xcite , is , in its turn , closely related to the ablowitz
ladik hierarchy ( alh ) @xcite .
so we shall establish the direct links between the model considered and the alh , which is much more well - studied than equations ( [ mainvec_eq ] ) , ( [ main_eq ] ) or models @xcite . in the present paper
we first derive the gauge equivalence between the heisenberg equation and the alh ( sections [ sec_heisalh ] , [ sec_gauge ] ) and then use it to obtain soliton solutions ( section [ sec_sol ] ) and the elliptical quasiperiodic ones ( section [ sec_qps ] ) .
the method used here , which may be called the ` embedding into the alh ' method , has been discussed in @xcite .
its main idea is that some equations can be , in some sense , ` derived ' from the system of differential - difference equations ( dde ) belonging to the alh , which means that any common solution of several equations from the alh also solves the equation we are dealing with .
relatively to the problem considered this can be briefly outlined as follows .
consider the system of two equations from the alh , @xmath26 where @xmath27 equation ( [ dnlse ] ) is the well - known discrete nonlinear schrdinger equation ( dnlse ) @xcite , modified by the substitution @xmath28 , while the next one , ( [ dmkdv ] ) , is the discrete modified kdv equation ( dmkdv ) @xcite .
these equations can be rewritten in terms of the complex variables @xmath29 , @xmath30 as @xmath31 where @xmath32 stands for @xmath33 and @xmath34 for @xmath35 .
it is very important that these equations are compatible , since they belong to the same hierarchy , and the constants of motion that play the role of the hamiltonians for the flows ( [ dnlse ] ) , ( [ dmkdv ] ) are in involution .
hence , we can consider them simultaneously , as one system of two equations .
it has been shown in @xcite , and one can easily verify this fact by simple calculations , that , for any fixed @xmath36 , each solution of system ( [ dnlse ] ) , ( [ dmkdv ] ) also solves the field equations of the @xmath24 nonlinear @xmath25-model , @xmath37 and that the quantities @xmath38 satisfy the 2d toda lattice equations @xmath39 ( see @xcite ) .
namely this we bear in mind when say that the o(3,1 ) nonlinear @xmath25-model and the 2d toda lattice can be ` embedded ' into the alh .
the situation with the heisenberg spin system is somewhat more difficult .
solution for equation ( [ main_eq ] ) can not be constructed by means of @xmath40 s and @xmath41 s only . to do that we have to consider the zcr for the alh and to analyze the corresponding linear problems .
the integrable ddes ( [ 1 ] ) , ( [ -1 ] ) , as well as all equations of the alh , can be presented as the compatibility condition for the linear system @xmath42 where @xmath43 and the matrices @xmath44 , @xmath45 are given by @xmath46 one can easily see that equations ( [ zcr - sp ] ) , ( [ zcr - evolp ] ) and ( [ zcr - sp ] ) , ( [ zcr - evoln ] ) are compatible only if matrices @xmath47 , @xmath48 and @xmath49 satisfy the so - called zero - curvature equations @xmath50 and @xmath51 which are equivalent to ( [ 1 ] ) and ( [ -1 ] ) correspondingly .
namely the solutions of the linear problems ( [ zcr - sp])([zcr - evoln ] ) , @xmath52 s , are the key objects of our consideration and the main result of this paper can be formulated as follows : for any @xmath36 , matrices @xmath53 constructed of solutions of the linear problems of the alh solve the matrix landau
lifshitz equation ( [ main_eq ] ) . to derive this result
consider matrices @xmath54 , @xmath55 defined by @xmath56 where @xmath21 is a pauli matrix ( [ pauli ] ) , and @xmath52 , recall , is a matrix solution of system ( [ zcr - sp ] ) , ( [ zcr - evoln ] ) ( in this notation @xmath57 ) .
it follows from ( [ sigma ] ) and ( [ zcr - evolp ] ) , ( [ zcr - evoln ] ) that @xmath58 \psi_{n } , \qquad \bar\partial \sigma^{a}_{n } = \psi_{n}^{-1 } \left [ \sigma^{a } , w_{n } \right ] \psi_{n}.\ ] ] using expressions ( [ vw ] ) for @xmath44 and @xmath45 one can find the derivatives of the matrices @xmath59 in terms of the matrices @xmath60 given by @xmath61 as follows : @xmath62 these relations together with analogous expressions for the derivatives @xmath63 , @xmath64 and formulae ( [ 1 ] ) , ( [ -1 ] ) , after straightforward calculations , omitted here , lead us to @xmath65 noting that @xmath66 and @xmath67 = 4i \left ( \lambda r_{n-1 } - \lambda^{-1 } r_{n } \right ) s_{n}^{+ } + 4i \left ( \lambda q_{n } - \lambda^{-1 } q_{n-1 } \right ) s_{n}^{-}\ ] ] ( both of these formulae follow from ( [ dsdz ] ) , ( [ dsdbz ] ) ) we obtain that for every @xmath36 the matrix @xmath59 solves the equation @xmath68 = 0\ ] ] which is the main equation of our study , ( see ( [ main_eq ] ) ) .
this key result of the present paper can be reformulated in terms of the vector @xmath69 , which corresponds to the matrix @xmath59 and which can be presented as @xmath70 where @xmath71 ( here @xmath72 are the elements of the matrix @xmath52 ) , as follows : for each @xmath36 the vector @xmath69 defined by ( [ vecsn ] ) , ( [ vecsn123 ] ) solves equation ( [ veczz ] ) .
thus we have established the links between equation ( [ main_eq ] ) , or ( [ veczz ] ) , describing stationary moving structures in the @xmath0-dimensional classical continuous heisenberg spin system and the alh .
some more detailed analysis of the gauge equivalence between these models one can find in the next section .
however , in this work we are going to focus our attention on ` practical ' aspects of this relation , so a reader can consider it as an ` empirical ' fact which can be straightforwardly , and rather easily , verified by the calculations outlined above . as was mentioned earlier , model ( [ tijon ] )
is known to be integrable and its zero - curvature representation has already been written out .
but , to our knowledge , the corresponding ist has not been elaborated yet , while the alh is one of the best - studied nonlinear integrable models . besides , the heisenberg equation is a vector problem , which somehow complicates inverse scattering analysis , while the alh is a scalar one .
so , to our opinion , the ` embedding into the alh ' approach is rather promising and in what follows we demonstrate its usefulness by constructing the soliton and quasiperiodic solutions for the equations considered using the already known solutions for the alh .
the magnetic energy density , @xmath73 , @xmath74 = \frac { g}{2 } \left ( \nabla \vec s , \nabla \vec s \right ) \label{w}\ ] ] of the field configurations obtained by the embedding into the alh method can be expressed in terms of the @xmath40 and @xmath41 s : @xmath75 = - \frac { v^{2}}{4 g } \left ( q_{n}r_{n } + q_{n-1}r_{n-1 } \right ) .
\label{wqr}\ ] ] it can be shown that from the viewpoint of application of solutions of the alh equations to the description of the vector field @xmath76 one has restrict himself with the case of @xmath77 , @xmath78 when the components of the vector @xmath69 ( [ vecsn ] ) are real ( in the opposite case , @xmath79 , the components of @xmath59 are complex ) and the magnetic energy ( [ wqr ] ) is positive . in the next section we will consider the relation between equation ( [ main_eq ] ) and the alh in the framework of the ist .
in the previous section we considered the relation between the alh and the landau lifshitz equation in terms of _ solutions _ : we demonstrated how to use solutions of the alh to obtain ones for the landau
lifshitz equation .
now we are going to discuss this question in somewhat more general way .
both the alh and the landau
lifshitz equations are integrable models and it is interesting to describe this correspondence in the language of the ist and to derive links between the auxiliary linear problems which are used to present the integrable models in the zero curvature form ( namely this is usually understood when one uses the words ` gauge equivalence ' ) . let us consider again the auxiliary linear problems of the alh mentioned in section [ sec_heisalh ] .
to our current purposes we do not need the discrete problem ( [ zcr - sp ] ) and will be dealing with the continuous ones ( [ zcr - evolp ] ) , ( [ zcr - evoln ] ) .
so , we omit now the index @xmath36 and rewrite ( [ zcr - evolp ] ) , ( [ zcr - evoln ] ) , ( [ vw ] ) as @xmath80 where @xmath81 and @xmath82 ( we have replaced @xmath40 , @xmath41 with @xmath83 , @xmath84 and @xmath85 , @xmath86 with @xmath87 , @xmath88 ) . in what follows we denote the spectral parameter by @xmath89 and use @xmath90 for its particular value appearing in the definition ( [ sigma ] ) of the matrix @xmath91 , @xmath92 the compatibility ( zero - curvature ) condition for the system ( [ zcr - evol - g ] ) @xmath93 = 0 \label{zcr - g}\ ] ] leads to the following system of four _ partial differential equations _ ( pde ) for four unknown functions @xmath83 , @xmath84 , @xmath87 , @xmath88 : @xmath94 this system is in some sense intermediate between the ddes ( [ 1 ] ) , ( [ -1 ] ) and the pde ( [ o31 ] ) : both of them can be ` reconstructed ' from ( [ q0-g])([r0-g ] ) ( we will return to this question below ) . and namely system ( [ q0-g])([r0-g ] ) is , strictly speaking , gauge equivalent to the spin field equation we are dealing with .
now we will derive the zcr for the stationary @xmath0-dimensional landau lifshitz equation from ( [ zcr - evol - g ] ) using the gauge transformation by means of the matrix @xmath95 . introducing the matrix function @xmath96 @xmath97
one can obtain from ( [ zcr - evol - g ] ) that it satisfies the following equations @xmath98 where @xmath99 \psi(\lambda ) , \\
w_{l}(\zeta ) = \psi^{-1}(\lambda ) \left [ w(\zeta ) - w(\lambda ) \right ] \psi(\lambda).\end{gathered}\ ] ] noting that @xmath100 \psi(\lambda ) = 2i \lambda^{-1 } \psi^{-1}(\lambda ) \left(\begin{array}{cc } 0 & -r_{0 } \\ q_{1 } & 0 \end{array}\right ) \psi(\lambda ) , \\
s_{\bar z}= \psi^{-1}(\lambda ) \left [ \sigma^{3 } , w(\lambda ) \right ] \psi(\lambda ) = 2i \lambda \psi^{-1}(\lambda ) \left(\begin{array}{cc } 0 & r_{1 } \\ -q_{0 } & 0 \end{array}\right ) \psi(\lambda)\end{gathered}\ ] ] one can present @xmath101 and @xmath102 as @xmath103 using the zero - curvature conditions for equations ( [ zcr - evol - ll ] ) , @xmath104 = 0,\ ] ] and calculating the left - hand - side part of this equation @xmath105 = \nonumber\\ \qquad{}= \frac{1}{2 } \left ( \frac{\lambda}{\zeta } - \frac{\zeta}{\lambda } \right ) \left\ { s s_{z \bar z } + \frac{1}{2 } \left ( s_{z } s_{\bar z } + s_{\bar z } s_{z } \right ) + i \lambda^{2 } s_{z } + i \lambda^{-2 } s_{\bar z } \right\}\end{gathered}\ ] ] one can conclude that the matrix @xmath91 must solve @xmath106 noting that the anticommutator of traceless @xmath107 matrices is proportional to the unit one and that the anticommutator of @xmath91 and @xmath108 or @xmath109 is zero ( which follows from the fact that @xmath110 , which is another form of the equality @xmath111 ) one can present this equation in the form ( [ main_eq ] ) . thus the linear problems ( [ zcr - evol - ll ] ) together with definitions ( [ v - ll ] ) and ( [ w - ll ] ) can be viewed as the zcr for the main equation of the present paper . after we have derived the zcr for ( [ main_eq ] ) we would like to make a few remarks on the application of the inverse scattering technique to non - evolutionary equations as ours .
the ist has been originally developed for the cauchy problems . however , since then much efforts has been made to adjust this method for various boundary value problems .
this is a rather difficult task since the latter seem to be more difficult than the former ones . among successes in this field
one should mention results related to the hyperbolic systems such as , e.g , the sine - gordon equation , the principal chiral field equations etc .
for these models the initial value
boundary value problems has been shown to be well stated problems , the existence and uniqueness of the solution has been established and ist - based algorithms to solve , say , the goursat type boundary problem have been elaborated .
one can find discussion of some recent results on boundary problems for the @xmath6-dimensional systems on semi - infinite and finite interval , for example in @xcite .
as to the elliptical systems , similar to the one discussed here , which do not possess characteristics , it is also possible to apply the ist for solving some boundary value problems .
usually it is achieved by breaking the symmetry between the coordinates ( which is , of course , not very natural for this kinds of equations ) , selecting one of them ( say , @xmath112 ) , considering the problem on a half - plane ( @xmath113 ) or on a finite domain ( in this case the part of the boundary data plays role of the cauchy conditions ) and performing analysis ( i.e. solving the direct and inverse spectral problems ) for the auxiliary linear equation corresponding to the complementary coordinate ( say , @xmath114 ) .
one can find examples of such approach in @xcite and references therein .
however , in this paper we do not discuss the mathematically rigorous formulation of the problem related to ( [ veczz ] ) .
we consider here the ist as a method to generate some classes of particular solutions and restrict ourselves to the ones most interesting from the physical viewpoint , solitons and quasiperiodic solutions . above
we have mapped the @xmath48-@xmath49 pair for system ( [ q0-g])([r0-g ] ) into the @xmath101-@xmath102 pair for equation ( [ main_eq ] ) by means of the gauge transformation ( [ phi - g ] ) , @xmath115 now we are going to derive the inverse transform : from ( [ zcr - evol - ll ] ) to ( [ zcr - evol - g ] ) ( i.e. from the @xmath48-@xmath49 pair ( [ v - ll ] ) , ( [ w - ll ] ) to ( [ v - alh ] ) , ( [ w - alh ] ) ) . the fist step is to diagonalize a solution of the landau
lifshitz equation , i.e. , to calculate , for given @xmath91 , the matrix @xmath116 defined by @xmath117 it is obvious that the @xmath118 correspondence is not one - to - one . for
any @xmath116 satisfying ( [ diag ] ) the matrix @xmath119 with an arbitrary diagonal matrix @xmath120 will also solve ( [ diag ] ) .
the main point of the landau
lifshitz equation @xmath121 alh transform is to use this arbitrariness to present the matrices @xmath122 , @xmath123 in ( [ v - alh ] ) , ( [ w - alh ] ) form with @xmath124 . @xmath125
this step needs some calculations which are presented in the appendix . performing then the gauge transform with the found matrix @xmath116
, one can obtain that the transformed @xmath101 , @xmath102 matrices @xmath126 are exactly of the form ( [ v - alh ] ) , ( [ w - alh ] ) which means that the functions @xmath87 , @xmath88 , @xmath83 , @xmath84 ( which are defined now in the terms of the matrix @xmath116 ( i.e. in the terms of the matrix @xmath91 ) solve the system ( [ q0-g])([r0-g ] ) . system ( [ q0-g])([r0-g ] ) that can be rewritten as the ddes from the alh .
indeed , starting from the quantities @xmath87 , @xmath88 , @xmath83 , @xmath84 one can _ define _ the quantities @xmath127 , @xmath128 @xmath129 and demonstrate that they satisfy the following identities : @xmath130 analogously , the quantities @xmath131 , @xmath132 , @xmath133 satisfy @xmath134 this procedure can be repeated in both directions @xmath135 this gives an infinite sequence of @xmath40 s , @xmath41 s which solve @xmath136 and @xmath137 i.e. the ablowitz ladik ddes . to conclude this section we want to discuss the following question .
if we start with the alh , which is a system of ddes , then the relation between the _ discrete _ equations ( alh ) and the _ partial differential _ landau
lifshitz equation is rather obvious : our pde is a differential consequence of the ddes .
but if we start with the landau lifshitz equation , then what role do the ddes from the alh play in the theory of our pde ? in simpler words , what does the subscript @xmath36 mean in terms of our pde ?
the answer is as follows .
we have an example of the situation studied by levi , benguria @xcite , shabat , yamilov @xcite and others : discrete integrable equations ( the equations from the alh in our case ) describe sequences of the bcklund transformations for some pdes ( the stationary @xmath0-dimensional landau lifshitz equation in our case )
. indeed , if we have a solution of our equation , @xmath138 , we can derive from it the matrix @xmath139 which solves the linear problems for the dnlse and dmkdv , and hence the quantities @xmath87 , @xmath88 , @xmath83 , @xmath84 which solve ( [ q0-g])([r0-g ] ) .
then we can construct the new @xmath116-matrix @xmath140 , and the new spin field @xmath141 which corresponds to the matrix @xmath142 .
this vector field will also solve the landau lifshitz equation .
this procedure can be repeated infinitely @xmath143 moreover , it can be performed in other direction @xmath144 thus we can obtain an infinite number of solutions ( , @xmath145 , @xmath146 , @xmath147 , @xmath141 , ) from one solution @xmath76 and relations between the vectors @xmath69 with different values of the index @xmath36 ( bcklund relations ) can be described by the equations which are analogous to ( and can be derived from ) the ddes from the alh .
the discrete nonlinear schrdinger equation ( [ dnlse ] ) under the condition ( [ kappa ] ) has been already solved in the pioneering work by ablowitz and ladik @xcite . as to the solutions of ( [ dmkdv ] ) , or system ( [ 1 ] ) , ( [ -1 ] )
, they can be obtained by minor modifications of the ones for ( [ dnlse ] ) , which is again a manifestation of the fact that all of them belong to the same hierarchy .
we will not repeat here the derivation of the ist ( one can find the technical details in @xcite or , say , in the book @xcite ) and write down only some final formulae that will be used below . the @xmath148-soliton solution of equations ( [ 1 ] ) , ( [ -1 ] ) can be presented as follows : @xmath149^{(jk ) } \lambda_{k}^{-1}.\ ] ] the constants @xmath150 s are the eigenvalues of the corresponding scattering problem ( [ zcr - sp ] ) ( to be more precise , the discrete spectrum of the scattering problem ( [ zcr - sp ] ) consists of @xmath148 pairs of the eigenvalues @xmath151 ) .
the functions @xmath152 are given by @xmath153 where @xmath154 s are arbitrary constants , @xmath155 while the matrix @xmath156 is given by @xmath157 here @xmath158 is the @xmath159 unit matrix , the overbar stands for the complex conjugation , @xmath160 and @xmath161 is the @xmath162 matrix with the elements @xmath163 solution for system ( [ zcr - sp])([zcr - evoln ] ) can be presented in the pure soliton case as @xmath164 where @xmath165 is the matrix of the following structure : @xmath166 here @xmath167 the above formulae contain all we need to construct solutions for the landau
lifshitz equation ( [ veczz ] ) , or ( [ main_eq ] ) .
the vertical component of the vector @xmath69 ( see ( [ vecsn123 ] ) ) can be presented , using ( [ psif])([psifg ] ) and the fact that in our case @xmath168 ( see remark after ( [ veczz ] ) ) , as @xmath169 while the horizontal components can be written as @xmath170 where @xmath171 the magnetic energy density ( [ w ] ) in this case can be presented , using ( [ wqr ] ) and the identity @xmath172 as @xmath173 = \frac { v^{2}}{4 g } \left\ { \frac { f_{n}(\infty ) } { f_{n+1}(\infty ) } + \frac { f_{n-1}(\infty)}{f_{n}(\infty ) } - 2 \right\}.\ ] ] noting that , for a fixed value of the index @xmath36 , the dependence on @xmath36 can be taken into account by the redefinition of the constants @xmath154 , we may chose @xmath174 and write the final formulae as follows : @xmath175 and @xmath176 where @xmath177 while the distribution of the magnetic energy of the field configuration given by ( [ s3])([s2 ] ) can be written as @xmath178 these formulae describe the @xmath148-soliton solutions of the stationary @xmath0-dimensional landau lifshitz equation . to make clear what kind of solutions we have obtained from the alh - solitons
let us consider in a more detailed way the simplest of the above solutions , namely the one - soliton ones . in this case
the quantity @xmath179 ( see ( [ anjk ] ) ) using the designation @xmath180 can be rewritten as @xmath181 where @xmath182 and @xmath183 , @xmath184 are some constants . setting @xmath174 , returning to the real coordinates @xmath185 , @xmath112 and @xmath186 , and introducing the vectors @xmath187 one can rewrite these formulae as @xmath188 with @xmath189 ( here braces stand for the usual scalar product : for @xmath190 , @xmath191 ) , from which one can derive the following expressions for the components of the vector @xmath76 .
the vertical components , @xmath192 , can be written as @xmath193 where @xmath194 , or , equivalently , as @xmath195 with the angle @xmath196 being given by @xmath197^{-1/2}. \label{sol_tan}\ ] ] the horizontal components @xmath198 can be presented as @xmath199 with @xmath200 here , the vector @xmath201 , @xmath202 is parallel to the velocity vector , @xmath203 , and is related to @xmath204 , @xmath205 by @xmath206 .
the function @xmath207 is given by @xmath208 the magnetic energy of this field configuration can be written as @xmath209 where @xmath210 stands for @xmath211 .
it can be shown that @xmath73 ( [ energy_1 ] ) is a second - order polynomial in @xmath192 : @xmath212 where @xmath213 is some constant . the linear energy density , @xmath214 , @xmath215 can be easily shown to be @xmath216 to simplify the following analysis let us consider the case when the velocity vector is directed along the @xmath185-axis ( @xmath217 , @xmath218 ) .
this does not lead to loss of generality because solutions corresponding the arbitrary vector @xmath219 can be obtained from the ones presented below by the substitution @xmath220 , @xmath221 .
it can be easily seen that formulae ( [ s3one])([varphione ] ) in the limiting cases @xmath222 and @xmath223 describe essentially different field structures . in the case
@xmath222 ( @xmath224 ) @xmath225 with an arbitrary constant @xmath226 , and both @xmath192 and @xmath73 depend on @xmath227 only , @xmath228^{1/2 } } \right),\\ { \cal w } = { \cal w } ( x - v t ) = g k_{\|}^{2 } \sin^{2 } \theta ( x - v t ) \end{gathered}\ ] ] while @xmath229 can be written as @xmath230 where @xmath231 so , this solution describes a localized structure , moving in the @xmath185-direction , which is phase modulated in the transversal direction ( @xmath112-direction ) , and it may be termed ` quasi one - dimensional soliton ' ( see fig . 1 ) . and @xmath232.,width=529 ] the soliton obtained above is essentially two - dimensional structure and despite apparent similarity it can not be reduced to its one - dimensional analogue .
indeed , in the one - dimensional case soliton solutions of equation ( [ heis11 ] ) possess the following form : @xmath233 and one can say that such solitons exist due to the _ temporal _ phase modulation of the whole medium , which manifests itself in the fact that @xmath234 , i.e. , soliton vanishes with @xmath235 going to zero .
in other words , these one - dimensional soliton structures do not exist in absence of the phase modulation @xmath236 . in our case ,
existence of solitons is due to the _ spatial _ phase modulation ( in @xmath112-direction ) , which manifests itself in the fact that the magnetic energy density is proportional to @xmath237 . in other words ,
solitons we have obtained differ from their one - dimensional analogues in the physical mechanism lying in their background : they are caused by the competition between the _ spatial _ dispersion and nonlinearity while in the one - dimensional case solitons are caused by the competition between the _ temporal _ dispersion and nonlinearity . in the opposite case
, @xmath238 ( @xmath239 ) , @xmath240 i.e. @xmath192 , which can be written as @xmath241 depends on @xmath112 only ( and , what is essential , does not depend on time ) , while the horizontal components are rotating with constant frequency : @xmath242 where @xmath243 and @xmath244 , remind , stands for @xmath245 .
this solution describes the spin wave localized in the @xmath246-neighborhood of the line @xmath247 ( this field distribution , which is depicted schematically in the fig . 1
, may be termed ` channel ' ) .
the magnetic energy of the ` channel ' field configuration does not depend on time , @xmath248 , hence it can be considered as almost static , in the sense that we have no energy transport in this case .
similar structures have been found by a s kovalev @xcite . and
@xmath249 , width=377 ] and @xmath250,width=377 ] the character of the soliton field structure in the general case @xmath251 can be seen from the fig . 2 .
the many - soliton structures in the general case can be viewed as consisting of several intersecting solitons .
one can find typical two - soliton spin distribution in the fig .
the alh in the quasiperiodic case is less studied than in the soliton one .
several authors have discussed the quasiperiodic solutions ( qps ) for the discrete nonlinear schrdinger and the discrete modified korteveg - de vries equations ( see , e.g. , @xcite ) , but their results are not enough to construct the corresponding solutions for the heisenberg equation using the ` embedding into the alh ' method . what we need and what is absent in the papers @xcite is a solution of the auxiliary system ( [ zcr - sp])([zcr - evoln ] ) which in the quasiperiodic case is known as the baker akhiezer function .
later this question was solved in @xcite ( see also @xcite ) .
however , these results , describing general finite - genus solutions , are rather cumbersome , so here we restrict ourselves only to the elliptic solutions , which are the simplest qps . the elliptic solutions for system ( [ 1 ] ) , ( [ -1 ] ) possess the following structure : @xmath252 where @xmath253 is one of the elliptic theta - functions ( see , e.g. @xcite ) , the phase @xmath254 is some linear function of the coordinates @xmath255 and @xmath256 ( it will be specified below ) , @xmath257 and the constants @xmath258 , @xmath259 are related by @xmath260 the quantities @xmath38 s can be presented as @xmath261 which follows from expressions ( [ qp_qr ] ) and the identity @xmath262 this identity is the fay s formulae @xcite for the elliptic functions .
it can be used to calculate the derivatives of the @xmath263-functions . differentiating ( [ fay ] ) with respect to @xmath255 and putting
@xmath264 one can obtain for the logarithmic derivative @xmath265 , @xmath266 the relation : @xmath267 using the latter one can obtain that functions ( [ qp_qr ] ) satisfy equations ( [ 1 ] ) , ( [ -1 ] ) provided the scales @xmath268 and @xmath269 are chosen as @xmath270 and the phase @xmath254 is given by @xmath271 \frac { z}{l } + i \left [ \psi(\beta ) - \psi(\beta-\alpha ) \right ] \frac { \bar z}{\tilde l } + \mbox{const}.\ ] ] the baker akhiezer function of our problem ( i.e. the quasiperiodic solution for system ( [ zcr - evolp ] ) ) , ( [ zcr - evoln ] ) can be written as a matrix with the elements @xmath272 here @xmath273 are arbitrary constants which are of no importance for our further consideration .
the phases @xmath274 are the linear functions of the coordinates , @xmath275 \frac { i z}{l } + \left [ \psi(\eta+\beta ) + \psi(\alpha ) - \psi(\alpha-\beta ) - \psi(\beta ) \right ] \frac { i \bar z}{\tilde l } , \\ \phi_{2 } = \left [ - \psi(\eta ) + \psi(\alpha+\beta ) - \psi(\alpha ) - \psi(\beta ) \right ] \frac { i z}{l } + \left [ - \psi(\eta + \alpha ) + \psi(\alpha ) \right ] \frac { i \bar z}{\tilde l } , \end{gathered}\ ] ] the quantities @xmath276 are given by @xmath277 and @xmath278 can be determined as a solution of the equation @xmath279 ( in the framework of the general theory , @xmath278 can be considered as the point of the riemann surface that corresponds to the point @xmath280 of the complex plane ) .
these formulae ( we do not present here the corresponding derivation procedure ) can be verified straightforwardly using ( [ psi_fay ] ) and ( [ fay ] ) .
they provide all necessary to construct the elliptic solutions for the heisenberg equation ( [ main_eq ] ) . using ( [ vecsn123 ] ) , ( [ theta ] ) and
( [ varphi ] ) , and omitting the @xmath36-dependence one can obtain @xmath281 and @xmath282 where @xmath283 the last two formulae can be rewritten as @xmath284 here the vector @xmath285 is given by @xmath286 with @xmath287 and the angle @xmath288 is defined by @xmath289 the magnetic energy density @xmath73 ( [ w ] ) of the above field configuration is , as in the one - soliton case , a second - order polynomial in @xmath192 : @xmath290 where @xmath213 is some constant . the last formula again illustrates the importance of the transversal modulation ( space dispersion ) for the existence of our nonlinear structures .
it should be noted that to ensure reality of all physical quantities , such as @xmath291 , @xmath73 one has to impose some restrictions on the parameters @xmath292 , @xmath293 and @xmath278 ( or @xmath294 ) which appear in the above expressions .
we can not at present formulate these restrictions in their general form , but will show below how these parameters should be chosen in some particular case , which is a generalization of the pure soliton one , in the sense that the one - soliton solutions obtained in section [ sec_sol ] are some limiting cases ot the elliptical ones discussed below .
thus , in what follows we restrict ourselves with the case of @xmath295 where @xmath296 is the complex half - period of the @xmath253-function ( see @xcite ) .
it can be shown that in this case both the components of the vector @xmath76 and the energy @xmath73 will be real if we choose @xmath297 where hats indicate that correspondent quantities are real . in what follows we use together with the theta - functions also the jacobian elliptical functions @xmath298 , @xmath299 and @xmath300 , @xmath301 with @xmath302 ( the definition of @xmath303 and analogous formulae for @xmath304 and @xmath305 one can find , e.g. , in @xcite ) .
the ` coordinate ' @xmath306 can be written now as @xmath307 where @xmath308 the expressions for @xmath196 and @xmath229 ( [ tg_theta_ell ] ) and ( [ varphi_ell ] ) , ( [ hatvarphi_ell ] ) can be presented as @xmath309 and @xmath310 . \label{ell_varphi}\ ] ] the vector @xmath311 is given by @xmath312 \vec k_{\bot},\ ] ] where @xmath313 while the vector @xmath314 is given by ( [ kappa_par_ell ] ) with @xmath315 and @xmath316 is related to @xmath317 and @xmath318 by @xmath319 it is straightforward to show that the limiting case of the elliptic quasiperiodic solutions presented above is solitons obtained in section [ sec_sol ] .
indeed , with the parameter @xmath320 ( [ ell_k ] ) going to zero ( which corresponds to @xmath321 ) , the elliptic functions sn , cn and dn become sin , cos and 1 correspondingly . noting that @xmath322 and identifying @xmath323 with @xmath324 ( which implies @xmath325 ) one can transform ( [ ell_tan ] ) and ( [ ell_varphi ] ) to formulae ( [ sol_tan ] ) and ( [ varphione ] ) describing solutions of the landau lifshitz equation in the one - soliton case .
to conclude , we want to summarize the main results and to outline some perspectives of the studies discussed in this paper . from the mathematical point of view , our main result is the established relation between the landau
lifshitz equation ( in the case @xmath326 ) and the alh . and
though we can not at present provide general explanation of what makes such apparently different models be so closely connected , we hope that the results presented in sections [ sec_sol ] , [ sec_qps ] are rather convincing arguments in favour of the fact that this relation is useful , at least from the practical standpoint , as a tool for generating of a large number of solutions . on the other hand , this work presents 2d stationary structures of the isotropic continuous heisenberg ferromagnetic spin system which have not been , to our knowledge , discussed in the literature and which seem to be interesting for the physics of magnetic phenomena .
it should be noted that we have obtained our results in the framework of the classical model , and one of the most important questions that should be solved now , from the viewpoint of applications to magnetism , is to develop quantum , or at least semi - classical , theory of such structures .
another question we want to mention here is the following one .
it is a widely known fact that solitons appear as a result of joint action of nonlinearity and some other mechanisms , such as dispersion . in our consideration
we have neglected the temporal dispersion ( temporal modulation ) , and its role has been played by the spatial one .
so , it is interesting to take into account both temporal and spatial dispersions , because the competition of different mechanisms in nonlinear regime can lead to nontrivial results . these and some other related questions may be the subject of further investigations .
this work was partly carried out during the authors stay at the abdus salam international centre for theoretical physics which is gratefully acknowledged for its kind hospitality and was partly supported by the ministerio de educacin , cultura y deporte of spain under grant sab2000 - 0256 .
we are grateful to a s kovalev for useful and stimulating discussions .
the aim of this section is to derive the matrix @xmath116 related to the solution of equation ( [ main_eq ] ) @xmath91 by @xmath327 such that the matrices @xmath48 , @xmath49 @xmath328 have the structure of the alh matrices ( [ v - alh ] ) , ( [ w - alh ] ) .
the diagonalization ( [ app : diag ] ) of a given matrix @xmath91 is not unique .
suppose we have found a matrix @xmath329 satisfying @xmath330 then any matrix @xmath331 with an arbitrary diagonal matrix @xmath332 satisfies ( [ app : diag ] ) .
hence , to solve our problem we can start from any solution of ( [ app : diag - f ] ) , for example from one given by @xmath333 ( one can verify by simple calculations that this is indeed a solution of ( [ app : diag - f ] ) ) and then to construct the diagonal matrix @xmath120 such that the matrices @xmath334 possess the properties we need . to simplify the following formulae
let us introduce the designation @xmath335 , @xmath336 and @xmath337 for the elements of the matrix @xmath91 , @xmath338 the main equation of this paper ( [ app : diag - f ] ) can be rewritten now as a system @xmath339 where @xmath340 and @xmath341 consider now the intermediate matrices @xmath342 , @xmath343 given by @xmath344 matrices @xmath48 ( [ app : v - psi ] ) and @xmath49 ( [ app : w - psi ] ) , which are the matrices @xmath345 , @xmath346 from section [ sec_gauge ] , can be written as @xmath347 and @xmath348 thus , if one takes the matrix @xmath120 such that its elements satisfy @xmath349 then matrices ( [ app : v ] ) , ( [ app : w ] ) shall have the following structure : @xmath350 as to other diagonal elements , they satisfy the identities @xmath351 using ( [ app : main_a])([app : main_c ] ) one can obtain that @xmath352 from the definition of the matrices @xmath48 , @xmath49 , @xmath342 , @xmath343 and @xmath91 it follows that @xmath353 and @xmath354 using again ( [ app : main_a])([app : main_c ] ) one can get that @xmath355 i.e. , comparing this result with ( [ app : ddiag ] ) , one can conclude that @xmath356 thus , setting @xmath357 , @xmath358 and defying the quantities @xmath87 , @xmath88 , @xmath83 , @xmath84 by @xmath359 one can rewrite matrices ( [ app : v ] ) , ( [ app : w ] ) as @xmath360 i.e. to present them in the form ( [ v - alh ] ) , ( [ w - alh ] ) . to summarize , we have derived , starting from a solution of the field equation ( [ main_eq ] ) , the matrix @xmath116 , defined by ( [ app : psi - df ] ) , ( [ app : f ] ) and ( [ app : dii ] ) , which can be used to perform the gauge transform from the landau lifshitz linear problems to the ones of the alh . 99 lakshmanan m , continuum spin system as an exactly solvable dynamical system , _ phys .
lett . _ * a61 * ( 1977 ) , 5354 .
takhtajan l a , integration of the continuum heisenberg spin chain through the inverse scattering method , _ phys .
_ * a64 * ( 1977 ) , 235237 .
kosevich a m , ivanov b a and kovalev a s , magnetic solitons , _ phys . rep .
_ , * 194 * ( 1990 ) , 117238 .
dodd r k , eilbeck j c , gibbon j. d and morris h c , solitons and nonlinear wave equations , academic press , london , 1984 .
l d and takhtajan l a , hamiltonian methods in the theory of solitons , springer - verlag , berlin , 1987 .
papanicolaou n , duality rotation for 2-d classical ferromagnets , _ phys .
_ * a84 * ( 1981 ) , 151154 .
vekslerchik v e , an @xmath24 nonlinear @xmath25-model and the ablowitz ladik hierarchy , _ j. phys . a : math .
* 27 * ( 1994 ) , 62996313 .
ishimori y , multi - vortex solutions of a two - dimensional nonlinear wave equation , _ progr .
* 72 * ( 1984 ) , 3337 .
lipovskii v d and shirokov a v , an example of gauge equivalence of multidimensional integrable equaitons , _ funk .
* 23 * ( 1989 ) , 6566 ( in russian ) .
ablowitz m j and ladik j f , nonlinear differential - difference equations , _ j. math .
* 16 * ( 1975 ) , 598603 .
vekslerchik v e , the 2d toda lattice and the ablowitz ladik hierarchy , _ inverse problems _ * 11 * ( 1995 ) , 463479 .
vekslerchik v e , the davey stewartson equation and the ablowitz ladik hierarchy , _ inverse problems _ * 12 * ( 1996 ) , 10571074 .
ablowitz m j and ladik j f , nonlinear differential - difference equations and fourier analysis , _ j. math
* 17 * ( 1976 ) , 10111018 .
ablowitz m j and segur h , solitons and the inverse scattering transform , siam , philadelphia , 1981 .
leon j and spire a , the zakharov shabat spectral problem on the semi - line : hilbert formulation and applications , _ j. phys . a : math .
* 34 * ( 2001 ) 73597380 ; nlin.ps/0105066 .
gutshabash e sh , lipovskij v d and nikulichev s s , nonlinear @xmath25-model in a curved space , gauge equivalence , and exact solutions of @xmath361-dimensional integrable equations , _ theor . math .
phys . _ * 115 * ( 1998 ) , 619638 ; translation from _ teor . mat .
_ * 115 * ( 1998 ) , 323348 ; nlin.si/0001012 .
levi d and benguria r , bcklund transformations and nonlinear differential - difference equations , _ proc .
* 77 * ( 1980 ) , 50255027 .
levi d , nonlinear differential - difference equations as bcklund transformations , _ j. phys . a : math .
* 14 * ( 1981 ) , 10831098 .
shabat a and yamilov r i , lattice representations of integrable systems , _ phys .
lett . _ * a130 * ( 1988 ) , 271275 .
kovalev a s , private communication .
ahmad s and chowdhury a r , on the quasiperiodic solutions to the discrete nonlinear schrdinger equation , _
j. phys .
a : math . gen . _ * 20 * ( 1987 ) , 293303 .
bogolyubov n n ( jr . ) , prikarpatskii a k and samoilenko v g , a discrete periodic problem for a modified nonlinear korteweg - de vries equation , _ teor . math .
* 50 * ( 1982 ) , 118126 ( in russian ) .
miller p d , ercolani n m , krichever i m and levermore c d , finite genus solutions to the ablowitz
ladik equations , _ comm . on pure and appl .
_ * 48 * ( 1996 ) , 13691440 .
vekslerchik v e , finite genus solutions for the ablowitz ladik fierarchy , _ j. phys . a : math .
* 32 * ( 1999 ) , 49834994 .
erdelyi a , magnus w , oberhettinger f and tricomi f g , higher transcendental functions , vol . 2 , mcgraw - hill book co. , new york , toronto , london , 1953 .
mumford d , tata lectures on theta i , ii , birkhauser , boston , 1983 , 1984 . | stationary structures in a classical isotropic two - dimensional continuous heisenberg ferromagnetic spin system are studied in the framework of the @xmath0-dimensional landau lifshitz model .
it is established that in the case of @xmath1 the landau
lifshitz equation is closely related to the ablowitz ladik hierarchy .
this relation is used to obtain soliton structures , which are shown to be caused by joint action of nonlinearity and spatial dispersion , contrary to the well - known one - dimensional solitons which exist due to competition of nonlinearity and temporal dispersion .
we also present elliptical quasiperiodic stationary solutions of the stationary @xmath0-dimensional landau lifshitz equation .
[ pritula - firstpage ] |
there have been collected many evidences ( due to polarization time - resolved experiments , @xcite computer simulations , @xcite and model studies @xcite ) which indicate that rotational and orientational relaxations of molecules in liquids under equilibrium and ( highly ) nonequilibrium conditions differ significantly .
the studies of nonequilibrium rotational relaxation have recently culminated in papers .
@xcite as has been demonstrated , relaxation of rotational energy of hot nonequilibrium photofragments slows significantly with the increase of their initial rotational temperature and differs dramatically from relaxation of the equilibrium rotational energy correlation function , manifesting thereby breakdown of the linear response description .
furthermore , hot nonequilibrium molecules exhibit almost dissipation - free coherent rotation , which persists for several rotational periods .
molecular dynamic simulations have attributed this unusual behavior to a rapid rearrangement of the local liquid structure : a hot solute pushes away the neighboring solvent molecules and rotates almost freely until it looses enough energy .
@xcite another way of thinking about that is in terms of the angular momentum dependent rotational friction : highly rotationally excited molecules experience lower friction than their thermally equilibrated counterparts .
@xcite these nonequilibrium relaxations are not reproduced by conventional theories of ( equilibrium ) molecular rotation , which are based upon master equations with constant relaxation rates ( the extended diffusion models , @xcite the keilson - storer model @xcite ) or rotational fokker - planck equations with constant frictions .
@xcite physically , this is not surprising .
if we consider relaxation under equilibrium conditions , then all the relevant energies are of the order of @xmath0 ( @xmath1 being the boltzmann constant and @xmath2 being the temperature of the bath ) .
so , putting aside non - markovian effects , we can always introduce a certain effective constant relaxation rate .
if we consider relaxation under nonequilibrium conditions , when initial energies are much higher than @xmath0 , such a description is no longer applicable , and the angular momentum dependence of relaxation rates must be explicitly taken into consideration ( see ref .
@xcite for more detailed argumentation and discussion ) .
there exist two major and complimentary approaches to rotational relaxation , via the fokker - planck equations @xcite and extended diffusion models .
@xcite in our previous paper , @xcite we have demonstrated how the angular - momentum dependent friction can be incorporated into the rotational fokker - planck equation .
the present paper deals with a similar generalization of the j - diffusion model .
first , we investigate how the angular momentum dependent rates manifest themselves in relaxations of initial nonequilibrium distributions and in behaviors of rotational and orientational correlation functions ( cfs ) .
second , we develop a reliable tool for the description of nonequilibrium molecular rotation , which can be used i.e. for the interpretation of photofragment anisotropy decays .
@xcite third , we demonstrate how the use of standard equilibrium
approaches may lead to quantitatively and even qualitatively wrong results , if they are applied to describe nonequilibrium molecular rotation .
photodissociation @xmath3 ( which was studied experimentally in the gas - phase @xcite and in the condensed phase @xcite , and also investigated via computer simulations @xcite ) is chosen in the present article as a prototype process which produces highly nonequilibrium hot photofragments .
the results of the molecular dynamics simulations @xcite are used to obtain realistic values of the model parameters and to test our theoretical predictions .
the structure of our paper is as follows . the generalized j - diffusion model , which accounts for the angular momentum dependence of the relaxation rate , is developed in sec .
a convenient explicit expression for the initial nonequilibrium angular momentum distribution is derived in section 3 .
relaxation of this nonequilibrium distribution , as well as dynamics of the angular momentum and energy cfs , are investigated in section 4 .
orientational relaxation is studied in section 5 .
the appendixes describe the methods of calculation of rotational cfs ( a ) , their relaxation times ( b ) , and orientational cfs ( c ) .
the reduced variables are used throughout the article : time , angular momentum and energy are measured in units of @xmath4 , @xmath5 and @xmath0 , respectively . here
@xmath1 is the boltzmann constant , @xmath2 is the temperature of the equilibrium bath , and @xmath6 is the moment of inertia of the photofragment , so that @xmath7 is the averaged period of its free rotation . for cn at @xmath8
k , @xmath9@xmath10 ps . all the laplace - transformed operators are denoted by tilde , viz . @xmath11
restricting our consideration to linear molecules and neglecting the short - time non - markovian effects , we start from the rotational master equation @xmath12 here @xmath13 is the probability density function , @xmath14 is the ( two - dimensional ) angular momentum of the linear top in its molecular frame , @xmath15 are the euler angles which specify orientation of the molecular frame with respect to the laboratory one .
the first term on the right - hand side of eq . ( [ kin1 ] ) is the free - rotor liouville operator , which describes the angular momentum driven reorientation , @xmath16 being the angular momentum operator in the molecular frame .
the last two terms are responsible for the bath - induced relaxation .
since our molecular ensemble is isotropic , the kernel of the dissipation operator , @xmath17 , is @xmath15-independent .
it must obey the detailed balance @xmath18 which is responsible for bringing the system under study to the equilibrium rotational boltzmann distribution@xmath19 normalization of eq .
( [ kin1]),@xmath20 which insures conservation of probability , is accounted for automatically . by selecting a particular form of the relaxation kernel @xmath21
, one can recover many models of rotational relaxation available in the literature : the extended diffusion models , @xcite the rotational fokker - planck equation , @xcite the keilson - storer model .
@xcite all these models , however , yield @xmath22 , that is the collision rate is angular momentum independent . for example
, the most popular model of rotational relaxation , the j - diffusion model , corresponds to the choice @xmath23 we can not straightforwardly generalize the j - diffusion model by replacing the rate @xmath24 with its @xmath25-dependent counterpart @xmath26 , since this procedure would violate normalization ( and therefore conservation ) of the probability density ( eq .
( [ norm ] ) ) . to get a correct description , we assume that the relaxation kernel can be represented in the factorized gaussian form @xmath27 @xmath28 , and @xmath29 being numerical constants to be specified . to satisfy the detailed balance ( [ detbal ] )
, we must choose @xmath30 then the generalized j - diffusion relaxation kernel takes the form @xmath31 here the rate @xmath24 governs the dissipation strength and the constant @xmath32 is chosen to insure that @xmath33 thus , @xmath34 can be regarded as @xmath25-dependent collision frequency
. @xcite eq .
( [ gjdkern ] ) looks as a natural generalization of the standard j - diffusion kernel ( [ jd ] ) , which still preserves its gaussian and factorized form .
if @xmath35 is small ( it is indeed small , see bellow ) , then the kernel ( [ gjdkern ] ) can be regarded as the first order correction to its j - diffusion counterpart ( [ jd ] ) .
furthermore , any relaxation kernel @xmath21 can be expanded in series on gaussians,@xmath36 @xmath37 being certain expansion coefficients .
@xcite thus , eq . ( [ gjdkern ] ) retains the first term in this expansion , and can further be refined by taking more terms , if necessary . the use of eq .
( [ gjdkern ] ) can be justified _ a posteriori , _ since the predictions of the generalized j - diffusion model in secs .
4 and 5 sustain comparison with the results of molecular dynamics simulations performed in refs . @xcite plugging the kernel ( [ gjdkern ] ) into eq .
( [ kin1 ] ) we obtain our generalized j - diffusion master equation : @xmath38 @xmath39 this is the equation which will be studied in the subsequent sections .
when * @xmath40 * ( @xmath41 ) , it reduces to the standard j - diffusion model . before embarking at particular calculations , it is useful to estimate plausible values of the parameters @xmath24 and @xmath35 .
since @xmath24 describes the overall dissipation strength and is similar to its counterpart in the standard j - diffusion model , there are no intrinsic limitations on the value of this parameter .
it is small ( @xmath42 ) for rarefied gases and large ( @xmath43 ) for liquids and solutions . as to the value of @xmath35
, it is expected to be small ( @xmath44 ) , since the effects due to the @xmath25-dependence of relaxation rates do not manifest themselves at room temperatures under equilibrium conditions . on the other hand ,
if the initial nonequilibrium energy of the photofragment @xmath45 is much larger than @xmath0 , then the product @xmath46 is not necessarily small .
it is in this case the effects due to the @xmath25-dependence of relaxation rates become important . note that the parameters @xmath24 and @xmath35 must be independent of initial conditions .
however , as we shall see in secs . 4 and 5 , eq .
( [ kin2 ] ) predicts that nonequilibrium responses , i.e. the ensuing rotational and orientational cfs , depend strongly upon initial conditions .
normally , master equations like ( [ kin2 ] ) are used to calculate various rotational and orientational cfs under equilibrium conditions , assuming that @xmath47 .
we wish to study rotational relaxation under nonequilibrium initial conditions @xmath48 to get an explicit expression for @xmath49 , we adopt a model developed in ref .
@xcite let us consider the photoreaction @xmath50 .
we assume that the photofragmentation is instantaneous on the time scale of molecular rotation .
this assumption holds true for most of small @xcite and polyatomic @xcite molecules , for which the photofragmentation time can be be estimated by few hundreds of femtoseconds .
@xcite having accepted the assumption about an instantaneous ( impulsive ) photodecomposition , we immediately conclude that dissociation produces a nonequilibrium distribution over the angular momenta of photoproducts .
indeed , there exist two major sources of rotational excitation of fragments .
these are the parent molecule rotation and the applied torque .
therefore , the angular momenta of the parent ( @xmath51 ) and product ( @xmath52 ) molecules are connected through the formula @xcite@xmath53 the first term describes mapping of the parent molecule rotation into that of the product , and the angular momentum @xmath54 originates from the impulsive torque arising due to the rupture of chemical bond(s ) of the parent .
the small latin indexes label the cartesian components of the corresponding vectors and tensors , @xmath55 and @xmath56 are the main moments of inertia of the species a and b , @xmath57 is the matrix of rotation from the frame of the main moments of inertia of the product to that of the parent , and @xmath58 are the pertinent euler angles .
let the parent molecule a be a planar asymmetric top ( i.e. , a triatomic molecule , @xmath59 ) and the product b be a linear rotor ( @xmath60 ) .
we then define the molecular frames of the parent ( @xmath61 ) and product ( @xmath62 ) molecules in such a way that the axes @xmath63 are perpendicular to the plane of the parent molecule , and @xmath64-axis coincides with that of the linear fragment ( see fig .
1 ) . in this case @xmath65
it is natural to surmise that the applied torque is perpendicular to the plane of the parent molecule , @xmath67 .
then , assuming that the parent molecules have a boltzmann equilibrium distribution at the bath temperature @xmath2 , we find that the linear photoproducts are distributed according to @xcite @xmath68^{2}+\xi^{2}j^{2}\sin^{2}\theta\right)\right\ } \label{nemy}\ ] ] ( hereafter , the subscript b , which denotes the angular momentum and moment of inertia of fragment b , is omitted and @xmath69 is the magnitude of the impulsive angular momentum @xmath54 ) .
we have introduced the parameters @xcite @xmath70 eq .
( [ parne ] ) is a desirable nonequilibrium distribution , which is employed in all the subsequent calculations . the entire information about dissociation is contained in the parameters @xmath69 , @xmath71 , and @xmath72 .
the first of them is responsible for the applied torque and can be considered as dynamical .
the last two parameters control the parent - product rotational energy transfer .
they are determined by the geometry of the dissociating molecule .
note that @xmath73 .
when @xmath74 , then the parent molecule is highly prolate ( @xmath75 ) and the linear fragment is " attached " perpendicularly to the parent molecule axis ( @xmath76 ) . if @xmath77 , then the parent molecule is also considerably prolate , but the linear fragment is " attached " parallel to the parent molecule axis ( @xmath78 ) . if the parent molecule is a symmetric top ( @xmath79 ) , then @xmath80 .
the quantities @xmath71 and @xmath72 can be regarded as known for a particular photoreaction , since the angle @xmath81 and the main moments of inertia are fixed for specific molecules a and b. the value of @xmath69 can be estimated through the consideration of the energy partitioning in the course of the photofragmentation .
@xcite distribution ( [ parne ] ) is convenient for the further use , since it covers many different scenario of the photofragmentation .
it reduces to the boltzmann equilibrium distribution ( [ boltz ] ) if @xmath82 and @xmath83@xmath80 . on the other hand ,
we get the delta - distribution @xmath84 in the limit @xmath85 .
the quantum version of distribution ( [ nemy ] ) is immediately written down after the `` quantization '' of the angular momentum : @xcite @xmath86 @xmath87^{2}+\xi^{2}\overline{j}^{2}\sin^{2}\theta\right)\right\ } .\ ] ] here @xmath88 is the rotational quantum number , @xmath89 eq .
( [ nemyq ] ) allows us to simulate a rotational distribution of any width centered at any value of @xmath88 .
several characteristic shapes of @xmath90 are presented in fig .
if @xmath80 , the distribution is symmetric . if @xmath91 ( for triatomics , this corresponds to dissociation from a bent configuration ) the distribution becomes asymmetric , with a long-@xmath88 bias .
let us now apply the above machinery to photodissociation @xmath3 .
this reaction proceeds via two channels , producing low energy and high energy photofragments .
the quantum distribution @xmath90 for nascent cn - fragments has been measured in .
@xcite a comparison of the `` full line '' distribution in fig .
2 with that reported in @xcite reveals that the former reproduces qualitatively the experimental results for hot photofragments . for cn at @xmath8
k , @xmath92 .
therefore , the `` realistic values '' of the dimensionless parameters @xmath93 , @xmath94 will be used in our illustrative calculations .
in this section , we study evolution of the quantities which depend on the angular momentum @xmath14 but are independent of the euler angles @xmath15 . we thus can integrate eq .
( [ kin2 ] ) over @xmath15 and arrive at the reduced master equation@xmath95 this equation can be used , for example , to calculate the angular momentum cf @xmath96 the averaged rotational energy@xmath97 and the rotational energy cf @xmath98 here @xmath99 and we use the notation @xmath100 , @xmath101 .
the corresponding integral relaxation times are determined as @xmath102 as is shown in appendix a , the laplace images of the probability density @xmath103 , as well as of cfs ( [ cj])-([cbe ] ) can be given in quadratures ( eqs .
( [ ronet ] ) , ( [ cab1 ] ) , and ( [ cabnewsf ] ) ) for any initial nonequilibrium distribution @xmath49 .
the inversion of the laplace transforms into the time domain can elementary be performed numerically .
furthermore , we derive elucidating analytical expressions in several important particular cases , which help us to grasp essential features of rotational relaxation under nonequilibrium conditions .
let us assume that the nonequilibrium distribution is given by eq .
( [ deltad ] ) , that is the photoproducts are produced with a fixed magnitude of the angular momentum @xmath69 .
such a situation corresponds to the procedure of the ( microcanonical ) preparation of hot molecules in simulations reported in @xcite .
we thus define the nonequilibrium photofragment energy @xmath104 we assume , in addition , that @xmath35 is small ( @xmath105 ) but the product @xmath106 can take any value . then eq .
( [ taud ] ) predicts that the cfs ( [ cj])-([cbe ] ) are all the same and exponential , @xmath107 the @xmath25-dependent rates ( [ z(j ) ] ) are seen to slow rotational relaxation .
the corresponding relaxation times ( [ tau ] ) read @xmath108 the last term in this expression is written in dimensional units , to emphasize that the generalized j - diffusion model predicts the arrhenius energy dependence of the rotational relaxation times .
( [ tauall ] ) corroborates the finding @xcite that there exist characteristic `` nonequilibrium energy '' @xmath109 and `` nonequilibrium temperature '' @xmath110 , @xmath111 starting from which the @xmath25-dependence of rotational relaxation becomes important .
if the characteristic nonequilibrium energy @xmath45 is much larger than @xmath0 , then the relaxation times for hot photofragments can substantially be longer than their equilibrium counterparts @xmath112 , despite @xmath35 is small .
the standard j - diffusion model , on the contrary , predicts that any rotational cf ( [ cab ] ) decays exponentially with the rate @xmath24 , irrespective of particular forms of @xmath113 , @xmath114 , and @xmath49 ( see eq .
( [ jdcf ] ) ) .
it is thus inadequate under highly nonequilibrium conditions , when @xmath115 .
the arrhenius energy dependence of the rotational relaxation times correlates with the results of tao and stratt , @xcite who simulated cfs @xmath116 ( [ sne ] ) for different initial nonequilibrium temperatures @xmath117 .
they obtained the following values : ( a ) @xmath118 for @xmath119 k , ( b ) @xmath120 for @xmath121 k , ( c ) @xmath122 for @xmath123 k ( see also ref .
we can take any two sets of @xmath124 and @xmath117 to calculate @xmath24 and @xmath35 by using eq .
( [ tauall ] ) . doing so for three different pairings of the sets , we have got remarkably consistent results : @xmath125 , @xmath126 for sets a&b , @xmath127 , @xmath128 for sets b&c , @xmath129 , @xmath126 for sets c&a . if we recall that eqs . ( [ taud1 ] ) and ( [ tauall ] )
are approximate and valid in the limit of @xmath130 , @xmath131 , the agreement becomes even more encouraging .
thus the quantity @xmath132 is close to @xmath133 ( @xmath134 ) at equilibrium ( @xmath135 k ) , @xmath136 at @xmath119 k , @xmath137 at @xmath121 k , and @xmath138 at @xmath123 .
so , rotational relaxation at @xmath123 k is nine times slower than that at the equilibrium temperature .
summarizing , the value of @xmath139 is a `` realistic value '' for the parameter which is responsible for the @xmath25-dependent relaxation .
it is timely to present an exact expression for the angular momentum integral relaxation time , which is valid for any values of the parameters of the generalized j - diffusion model . according to the formulas derived in appendix b , @xmath140 ( @xmath141 is defined via eq .
( [ vb ] ) ) .
if we assume that @xmath105 , then eq . ( [ tjexact ] ) reduces to ( [ tauall ] ) for any @xmath142 , and @xmath69 .
this lends an additional support to the validity of eqs .
( [ tauall ] ) . on the other hand , eq .
( [ tjexact ] ) demonstrates that the @xmath25-dependence of rotational relaxation rates works both ways : if @xmath69 is small and @xmath143 , then nonequilibrium integral relaxation times can be smaller then their equilibrium counterparts .
this is illustrated by fig .
3 , which shows the reduced angular momentum relaxation time @xmath144 ( full line ) and rotational energy relaxation times @xmath145 ( dashed line ) vs. parameter @xmath71 .
the nonequilibrium relaxation times are calculated via eqs .
( [ gen2])-([wb ] ) , and their equilibrium counterparts are given by eqs .
( [ taueq ] ) . the observed decrease of the relaxation times at @xmath146 and @xmath143 is not so pronounced as their exponential increase at @xmath147 .
4 compares rotational cfs ( [ cj])-([cbe ] ) calculated within the standard ( @xmath40 ) and generalized ( @xmath139 ) j - diffusion models .
the cfs have been computed as explained in appendix a. the realistic values of the model parameters ( @xmath148 , @xmath80 , @xmath149 , and @xmath150 ) have been used .
the standard j - diffusion predicts all the cfs to coincide with @xmath151 .
according to the generalized j - diffusion model , the cfs relax much slowly , as expected .
for the model parameters chosen , @xmath116 and @xmath152 are almost indistinguishable , @xcite while @xmath153 decays slower than the former two cfs .
if we assume that the initial distribution is given by delta - function ( [ deltad ] ) and @xmath105 , then the general eqs . ( [ ronet ] ) and ( [ sigd ] )
simplify to yield the following evolution of the probability density function : @xmath154 in an accord with what has been observed in , @xcite the rotational distribution ( [ ronet1 ] ) is clearly bimodal . at every time moment , @xmath103 is a mixture of the initial nonequilibrium distribution ( [ deltad ] ) and the equilibrium distribution ( [ boltz ] ) . the standard j - diffusion model predicts the same formula but with @xmath155 .
thus , the @xmath25-dependent rate effectively slows the rotational relaxation and increases the lifetime of the initial nonequilibrium contribution .
the higher is the value of @xmath106 , the stronger is the effect .
for example , if we take the dimensionless values of the parameters @xmath127 , @xmath128 obtained above , we can calculate the decay rates at @xmath123 k. namely , the generalized j - diffusion model predicts @xmath156 @xmath157 , while the j - diffusion model yields @xmath158 @xmath157 . the generalized j - diffusion rate explains why @xmath103 simulated for cn fragments at @xmath123 k shows both the nonequilibrium and equilibrium contributions for more than @xmath159 ps .
@xcite the standard j - diffusion model , which predicts the decay rate to be @xmath160 times higher , is thus absolutely inadequate far from equilibrium .
orientational cf of the rank @xmath88 is defined through the wigner d - functions @xcite as follows : @xmath161 its laplace image , @xmath162 , can be calculated as explained in appendix c , eq .
( [ ocf_fin ] ) . if we take the initial delta - function distribution ( [ deltad ] ) , then eq . ( [ ocf_fin ] ) can considerably be simplified . at short times ( @xmath163 ) , it can be inverted into the time domain to yield @xmath164 the numerical coefficients @xmath165 are given by eq .
( [ free ] ) .
the standard j - diffusion predicts the same formula but with @xmath155 .
thus , the oscillations in orientational cfs of hot photofragments , which have been measured in the gas phase @xcite and simulated in the condensed phase , @xcite are entirely determined by the initial nonequilibrium distribution ( [ deltad ] ) .
the period of these oscillations is uniquely determined by the value of @xmath69 .
5 compares orientational cfs simulated for hot cn fragments injected at the temperature @xmath123 k into the heat bath of argon atoms at @xmath166 k ( refs .
@xcite ) and those calculated within the generalized j - diffusion model .
the delta - function initial distribution ( [ deltad ] ) has been used in our calculations , since it corresponds to the procedure of the preparation of the ensemble of photofragments in the simulations .
the simulated orientational cfs look qualitatively very similar to those described by eq .
( [ ocfgg2 ] ) .
quantitatively , they can be fitted by eq .
( [ ocfgg2 ] ) relatively well , but the so - obtained decay rate of the second rank orientational cf turns out to be @xmath167 higher than that of the first rank cf .
the orientational cfs which are exactly calculated within the generalized j - diffusion model are seen to reproduce the simulated cfs very well .
this is remarkable , since we did not fit the simulated cfs : we used the parameters of the generalized j - diffusion model obtained in the previous section through the comparison of the simulated and theoretical energy relaxation times .
namely , we took @xmath128 , @xmath127 , and @xmath168 .
@xcite this lends an additional support to self - consistency and predictive strength of the generalized j - diffusion model .
( [ ocfgg2 ] ) makes it clear that the persistence of the oscillatory behavior for orientational cfs is much higher in the generalized j - diffusion model , due to the @xmath69-induced decrease of the relaxation rate , @xmath169 .
this is vividly illustrated by fig .
6 , in which presented are the first and second rank orientational cfs calculated within the generalized ( @xmath139 ) and standard ( @xmath40 ) j - diffusion models for the `` realistic values '' of the model parameters ( @xmath148 , @xmath80 , @xmath149 , and ) and @xmath170 .
the generalized j - diffusion model predicts the highly oscillatory orientational cfs . in exactly the same situation ,
the standard j - diffusion model predicts overdamped and slowly decaying cfs , which have nothing in common with the `` true '' cfs .
thus , the use of the standard ( equilibrium ) models of rotational relaxation beyond their domain of validity may lead to completely wrong predictions .
we can use eq .
( [ ocf_fin ] ) to derive a simple expression for orientational cfs in the case of strong dissipation , @xmath171 .
as expected , they are described by the diffusion formula @xmath172 with the diffusion coefficient @xmath173 equaled to the _ equilibrium _ angular momentum integral relaxation time , @xmath174 .
the latter is given by eq .
( [ taueq ] ) , so that @xmath175 the orientational relaxation times are calculated as @xmath176 it is popular to plot these quantities vs. the angular momentum relaxation times or rates .
@xcite that is why figs .
7 display the graphs @xmath177 ( a ) and @xmath178 ( b ) vs. @xmath24 calculated within the standard and generalized j - diffusion models . to get a better idea about the behavior of these curves
, it is insightful to obtain explicit expressions for @xmath179 in case of weak and strong dissipation . if @xmath180 , then eq . ( [ ocf_fin ] )
predicts that @xmath179 is inversely proportional to the rate @xmath24:@xmath181 here @xmath182 and @xmath183 are explicitly given by eq .
( [ free ] ) .
@xcite the standard j - diffusion formula @xcite is recovered at @xmath40 . again , if the product @xmath106 is @xmath184 or larger , then nonequilibrium initial conditions manifest themselves in a considerable increase of @xmath178 in comparison with the standard j - diffusion predictions at @xmath185 ( figs .
7b ) . when @xmath186 , the situation changes and the standard j - diffusion model overestimates the actual values of @xmath179 ( figs .
if @xmath171 then the diffusion formula ( [ dif ] ) holds and , therefore,@xmath187 this expression can be termed as the generalized hubbard relation . the standard hubbard relation , @xmath188^{-1}$ ] ,
@xcite is seen to be significantly off .
this discrepancy is remarkable , since it embodies the breakdown of the linear response theory .
indeed , the onset of rotational diffusion occurs after the molecular angular momenta have been thermolized according to the bath - induced boltzmann distribution ( [ boltz ] ) .
thus the rotational diffusion equation ( [ dif ] ) is a legitimate description at @xmath189 and the rotational diffusion coefficient is determined by the _ equilibrium _
rotational fluctuations and therefore by the _
value of the angular momentum relaxation time , @xmath174 . as has been detailed in section 4 , if the system starts far from equilibrium , then its angular momentum relaxation time @xmath190 can differ dramatically from its equilibrium counterpart , @xmath174 .
it is in this case we expect the breakdown of the linear response green - kubo - type formulas for transport coefficients , which identify the rotational diffusion coefficient @xmath173 with the angular momentum relaxation time @xmath190 .
an adequate description of rotational and translational relaxation in liquids under nonequilibrium conditions ( or within an interval of characteristic energies which highly exceeds that of the bath thermal energies ) can not be carried out in terms of conventional langevin and fokker - planck equations with constant frictions or by master equations with constant relaxation rates : one has to take into account ( linear and/or angular ) velocity dependence of friction .
@xcite the present paper deals with studying nonequilibrium rotational and orientational relaxation .
we have generalized the standard j - diffusion model by allowing its dissipation rate to be angular momentum dependent and calculated various nonequilibrium rotational and orientational cfs .
the reaction @xmath3 has been selected as a prototype process which produces highly nonequilibrium hot photofragments .
we have used the results of computer simulations @xcite to obtain realistic values of the model parameters and to test our theoretical predictions .
we have found that nonequilibrium rotational relaxation rates assume the form @xmath191 , where @xmath24 is the equilibrium rate , @xmath45 is the initial `` nonequilibrium energy '' , @xmath2 is the equilibrium temperature of the heat bath , and @xmath35 is the dimensionless small parameter fixed for a system under study .
accordingly , the relaxation times have the arrhenius energy dependence , @xmath192 .
so , there exist characteristic `` nonequilibrium energy '' @xmath193 and `` nonequilibrium temperature '' @xmath194 , starting from which the @xmath25-dependence of relaxation rate becomes important and induces a significant slowing of rotational relaxation . in agreement with simulations , @xcite relaxation of the rotational probability density
is shown to be a bimodal time - dependent mixture of the initial nonequilibrium and final equilibrium distributions .
the slowing of rotational relaxation induces a long lifetime of the initial nonequilibrium contribution , which manifests itself in pronounced coherent effects .
the slowing of rotational relaxation causes qualitative changes in molecular reorientation . in agreement with simulations ,
@xcite hot nonequilibrium molecules exhibit slightly perturbed coherent rotation , which persists for several rotational periods .
a similar observation has recently been made in the context of molecular excitation by strong femtosecond pulses : nonequilibrium molecular wave packets can be formed in such a way as to slow their subsequent rotational relaxation .
@xcite we have demonstrated that orientational cfs in the overdamped limit are described by the diffusion equation with the diffusion coefficient equaled to the _ equilibrium _ angular momentum integral relaxation time , @xmath174 .
thus the orientational relaxation times @xmath179 obey the generalized hubbard relation , @xmath195^{-1}$ ] . since the nonequilibrium angular momentum relaxation time @xmath190 can differ substantially from its equilibrium counterpart @xmath174 ,
the standard hubbard relation , @xcite in which @xmath174 is replaced by @xmath190 , can be significantly off .
it should be emphasized that all rotational and orientational cfs and their relaxation times dependent explicitly on the nonequilibrium preparation of the molecular ensemble .
this effect is unreproducible within the standard ( equilibrium ) rotational models , which predict that relaxation rates are independent of initial conditions and are given by the linear response theory .
thus , the arrhenius forms of the rotational relaxation times , slowing down of rotational relaxation , and violation of the hubbard relations are all manifestations of the breakdown of the linear response theory far from equilibrium . in practical terms ,
the standard and generalized j - diffusion models are almost indistinguishable under equilibrium conditions . under nonequilibrium conditions ,
their predictions differ dramatically .
the differences are not only quantitative but , not infrequently , qualitative .
the message is thus as follows : the friction and/or the relaxation rate must be taken @xmath25-dependent in order to adequately describe rotational relaxation far from equilibrium .
note , finally , that our theory can straightforwardly be generalized to symmetric and asymmetric top molecules . in this latter case
, the relaxation rate can be taken rotational energy dependent , rather than angular momentum dependent .
the method of solution of the corresponding kinetic equations remains absolutely the same , and the explicit formulas for various cfs can be written down after a straightforward generalization of the results presented in the appendixes .
our theory is also readily extendable to quantum case , by replacing the integrations over the angular momentum @xmath25 with summations over the rotational quantum number @xmath88 , and switching from the classical orientational cfs ( [ free ] ) to their quantum mechanical counterparts ( see , e.g. , ref .
such a generalization might be useful for describing rotational coherences and time - dependent alignments in nonequilibrium dissipative systems , like those studied in ref .
let @xmath113 and @xmath114 be arbitrary functions of the angular momentum
. then the cf@xmath196 can be evaluated through the rotational kinetic equation ( [ kinj ] ) as follows:@xmath197 after applying the laplace transform ( [ lap ] ) to eq .
( [ kinj ] ) with initial condition ( [ kinjt0 ] ) , we obtain@xmath198 here @xmath199 multiplying eq .
( [ kinjt0 ] ) by @xmath200 and integrating over @xmath14 , we obtain an algebraic equation for @xmath201 . inserting its solution into eq .
( [ kinjl1 ] ) , multiplying the so obtained expression by @xmath113 and integrating over @xmath14 , we get@xmath202 here we have introduced the functionals @xmath203 and their laplace transforms @xmath204 which are defined for any function @xmath205 . in principle , eq . ( [ cab1 ] ) delivers the desirable expression for cf @xmath206 which , after the numerical inversion of the laplace transform , yields the cf in the time domain , @xmath207 .
the long - time limit of this cf is determined as follows : @xmath208 here the averages are defined as @xmath209 if both @xmath210 and @xmath211 are nonzero , then cf ( [ cab ] ) possesses a stationary long - time asymptote ( [ cabinf ] ) .
this is so for the rotational energy cf , for example .
therefore , @xmath212 when @xmath213 .
this is undesirable for doing numerics .
it is more convenient to subtract this constant contribution and redefine the cf ( [ cab ] ) as follows:@xmath214 evidently , this cf is normalized to unity ( @xmath215 ) and does not have any stationary asymptote ( @xmath216 when @xmath217 ) . taking the laplace transform of eq .
( [ cabnew ] ) , we get@xmath218 eq . ( [ cabnews ] ) possesses yet an undesirable property : its numerator is a difference of two terms @xmath219 . of course , these two terms cancel each other when @xmath213 but it is convenient for numerical implementations to explicitly extract this singular contribution out of @xmath206 . to this end , we make use of the identity @xmath220 which is immediately derived from eq .
( [ psi1 ] ) . applying this formula several times ,
we can rewrite eq .
( [ cabnews ] ) in the following equivalent form@xmath221 here@xmath222 eqs .
( [ cabnewsf ] ) and ( [ ps ] ) are our final singularity - free formulas , which are used for the numerical inversion of the laplace transforms . the explicit formulas for cfs ( [ cj])-([cbe ] )
can be obtained by inserting the corresponding functions @xmath113 and @xmath114 into eqs .
( [ cabnewsf ] ) and ( [ ps ] ) .
for example , the angular momentum cf ( [ cj ] ) corresponds to @xmath223 . due to the isotropy of rotational space , @xmath224 , so that the general expression ( [ cabnewsf ] ) simplifies considerably : @xmath225 there is no such a simplification for cfs ( [ sne ] ) and ( [ cbe ] ) , and all the terms ( [ ps ] ) contribute into eq .
( [ cabnewsf ] ) .
we do not give the corresponding expressions explicitly , since they add nothing profound .
cf ( [ cabnew ] ) can be used to define an important quantity , the integral relaxation time:@xmath226 since @xmath227 , eqs .
( [ cabnewsf ] ) and ( [ ps ] ) predict that @xmath228 here @xmath229 if we assume that the relaxation rate is constant ( @xmath230 @xmath231 ) then the j - diffusion model is recovered . in this case , evidently , @xmath232 then , eqs . ( [ cabnewsf ] ) and ( [ tab ] )
simplify to @xmath233 irrespective of particular forms of @xmath113 , @xmath114 , and @xmath49 .
thus all rotational cfs in the j - diffusion model decay exponentially with the rate @xmath24 , and their integral relaxation times are all equaled to the inverse value of this rate .
if the relaxation rate is @xmath25-dependent ( @xmath234 ) , the situation is much more complicated .
however , the formulas simplify dramatically if the nonequilibrium distribution is given by the delta - function ( [ deltad ] ) . if we assume , additionally , that @xmath130 but the product @xmath106 can take any value , we can write then @xmath235 after the insertion of these expressions into eqs .
( [ cabnewsf ] ) and ( [ ps ] ) , we obtain @xmath236 for any @xmath113 and @xmath114 .
if we wish to follow how the initial nonequilibrium distribution @xmath49 transforms in time into the equilibrium one , @xmath237 , we must solve kinetic equation ( [ kinj ] ) with the initial condition @xmath238 .
the solution is given by a slightly modified version of eqs .
( [ kinjl1 ] ) and ( [ cab1]):@xmath239 @xmath240
if we employ the nonequilibrium initial distribution ( [ nemy ] ) , then the quantities @xmath241 ( eq . ( [ psi1 ] ) ) can be evaluated analytically for any cf of interest in the present paper ( @xmath242 ) .
this means that the integral relaxation times @xmath243 can also be calculated analytically . to this end
, convenient is to introduce the generating function@xmath244 @xmath245 evidently , @xmath246 @xmath247 eq . ( [ gen ] )
is elementary evaluated to yield@xmath248 differentiating this expression with respect to @xmath35 , we obtain:@xmath249 @xmath250 @xmath251 here @xmath252 @xmath253 these expressions can be substituted into eq .
( [ tab ] ) to calculate various integral relaxation times .
@xcite in general , the so - obtained expressions are quite cumbersome and are not presented here .
however , they are useful for obtaining simple elucidating formulas in several particular cases . if @xmath105 and @xmath147 , then @xmath254 if the initial distribution ( [ nemy ] ) reduces to the equilibrium one ( @xmath255 , @xmath82 ) , we get @xmath256 and @xcite@xmath257 it has not escaped our notice that the integral relaxation times @xmath243 ( [ tab ] ) diverge for @xmath258 , as predicted by eqs .
( [ gen2])-([wb ] ) .
although such situation is far beyond the expected domain of validity of our model ( @xmath259 ) , there is nothing pathological in the occurrence of the `` phase transition '' at @xmath260 and all cfs are well behaved for @xmath258 .
this means simply that the @xmath25-dependence of relaxation rates slows rotational cfs so dramatically that their integral relaxation times do not exist .
having inserted the definition of orientational cf ( [ ocf ] ) into eq .
( [ kin2 ] ) , we obtain the equation @xmath261 @xmath262 here @xmath263 are the matrix elements of the angular momentum operators @xmath264 over the d - functions : @xcite orientational cfs are calculated very similarly to rotational cfs ( see appendix a ) . first we introduce the free linear rotor orientational cf @xcite @xmath268 @xmath269 being the reduced wigner function . @xcite
we further define the functional and its laplace transform@xmath271 @xmath272 then , closely following the derivation of eq .
( [ cab1 ] ) , we obtain the following expression for the laplace transform of the orientational cf : @xmath273 since orientational cfs of dissipative molecules , as distinct from their bath - free counterparts ( [ free ] ) and rotational cfs ( appendix a ) , do not possess stationary asymptotes , this formula is well behaved in the limit @xmath213 and can be used for the numerical inversion of the laplace images into the time domain .
20 u. banin and s. ruhman , j. chem .
98 , 4391 ( 1993 ) .
e. lenderink , k. duppen , and d. a. wiersma , chem .
211 , 503 ( 1993 ) .
e. lenderink , k. duppen , f. p. x. everdij , j. marvi , r. torre , and d. a. wiersma , j. phys .
100 , 7822 ( 1996 ) .
c. wan , m. gupta , and a. zewail , chem .
256 , 279 ( 1996 ) .
s. gnanakaran , m. lim , n. pugliano , m. volk , and r. m. hochstrasser , j. phys . : condense matter 8 , 9201 ( 1996 )
. m. volk , s. gnanakaran , e. gooding , y. kholodenko , n. pugliano , and r. m. hochstrasser , j. phys .
a. 101 , 638 ( 1997 ) .
m. lim , s. gnanakaran , and r. m. hochstrasser , j. chem .
106 , 3485 ( 1997 ) .
t. khne and p. vhringer , j. phys .
a 102 , 4177 ( 1998 ) .
s. hess , h. brsing , and p. vhringer , j. chem .
111 , 5461 ( 1999 ) .
m. volk , j. phys . chem . a 103 , 5621 ( 1999 ) .
s. hess , h. hippler , t. khne , and p. vhringer , j. phys .
a 103 , 5622 ( 1999 ) .
h. brsing , j. lindner , s. hess , and p. vhringer , appl .
b. 71 , 411 ( 2000 ) .
h. brsing and p. vhringer , phys .
phys . 2 , 73 ( 2000 ) . h. fidder , f. tschirschwitz , o. duhr , and e. t. j. nibbering , j. chem .
114 , 6781 ( 2001 ) .
a. c. moskun and s. e. bradforth , j. chem .
119 , 4500 ( 2003 ) .
i. benjamin and k. r. wilson , j. chem .
90 , 4176 ( 1989 ) .
i. i. benjamin , u. banin and s. ruhman , j. chem . phys .
98 , 8337 ( 1993 ) .
i. benjamin , j. chem .
103 , 2459 ( 1995 ) .
n. winter , i. chorny , j. vieceli and i. benjamin , j. chem .
119 , 2127 ( 2003 ) .
a. i. krylov and b. b. gerber , j. chem .
100 , 4242 ( 1994 ) .
w. coffey , m. evans and p. grigolini .
molecular diffusion and spectra ( john wiley & sons , 1984 ) .
chapters 7 - 9 .
a. p. blokhin and m. f. gelin , chem .
252 , 323 ( 2000 ) .
m. f. gelin , j. mol .
93 , 51 ( 2001 ) .
a. s. moskun , a. e. jailaubekov , s. e. bradforth , g. tao , and r. m. stratt , science 311 , 1907 ( 2006 ) . g. tao and r. m. stratt , j. chem . phys . 125 , 114501 ( 2006 ) .
m. f. gelin and d. s. kosov , j. chem .
125 , 224502 ( 2006 ) .
r. g. gordon , j. chem .
44 , 1830 ( 1966 ) .
m. fixman and k. rider , j. chem .
51 , 2425 ( 1969 ) .
r. e. d. mcclung , adv .
10 , 83 ( 1977 ) .
a. i. burshtein and s. i. temkin .
spectroscopy of molecular rotation in gases and liquids ( cambridge university press , cambridge , 1994 ) .
r. a. sack , proc .
b. 70 , 402 ( 1957 ) ; ibid 70 , 414 ( 1957 ) .
m. f. gelin and d. s. kosov , j. chem .
124 , 144514 ( 2006 ) .
p. s. hubbard , phys .
a. 6 , 2421 ( 1972 ) .
g. w. ford , j. t. lewis and j. mcconnell , phys .
a. 19 , 907 ( 1979 ) .
a. morita , j. chem .
76 , 3198 ( 1982 ) .
d. h. lee and r. e. d. mcclung , chem . phys .
112 , 23 ( 1987 ) .
j. s. baskin and a. zewail , j. phys .
98 , 3337 ( 1994 ) .
j. s. baskin and a. h. zewail , j. phys .
a 105 , 3680 ( 2001 ) .
i. nadler , d. mahgerefteh , h. reisler , and c. wittig , j. chem .
82 , 3885 ( 1985 ) .
angular momentum dependent relaxation rates can be calculated for collisions of rigid nonspherical bodies , @xcite but the so - obtained formulas are very cumbersome and the very use of the predictions of the gas - phase ( uncorrelated binary collision ) theories is questionable for our purposes .
m. r. hoare , adv .
20 , 135 ( 1971 ) .
k. l. rider and m. fixman , j. chem .
57 , 2548 ( 1972 ) .
m. p. allen , g. t. evans , d. frenkel , b. m. mulder , adv . chem .
83 , 89 ( 1993 ) .
gaussian expansions are commonly employed in quantum chemical calculations .
we can also recall a similar factorized decomposition of the spectral density function in quantum relaxation theory .
@xcite c. meier and d. j. tannor , j. chem .
111 , 3365 ( 1999 ) .
a. p. blokhin and m. f. gelin , phys .
phys . 4 , 3356 ( 2002 ) .
a. p. blokhin , m. f. gelin , i. i. kalosha , s. a. polubisok and v. a. tolkachev , j. chem .
110 , 978 ( 1999 ) . in some cases
, the effect of rotational predissociation on the anisotropy evolution can be taken into account as proposed in ref .
@xcite note that the present parameter @xmath71 differs from its counterpart from ref .
@xcite by the factor of @xmath274 .
see discussion in supplementary material to ref .
@xcite for a validity of using the gas phase distribution for the interpretation of condensed phase data .
note that nonequilibrium cfs @xmath116 and @xmath152 simulated in @xcite are also very similar .
d. a. varshalovich , a. n. moskalev and v. k. hersonski .
quantum theory of angular momentum ( world scientific , singapore , 1989 ) . since the c -
n interatomic length has been taken as @xmath275 angstroms in , @xcite the free rotation period @xmath276 for cn at @xmath277 k is @xmath278 ps .
if we recalculate the nonequilibrium temperature @xmath123 into the dimensionless transferred angular momentum , we get then @xmath279 . we used a slightly smaller vale of @xmath168 in our generalized j - diffusion calculations presented in fig . 5 , in order to better reproduce the oscillation period of the cfs .
j. g. powles and g. rickayzen , mol .
33 , 1207 ( 1977 ) .
note that @xmath280 for odd @xmath88 .
thus , if @xmath281 , then also @xmath282 for odd @xmath88 . n. g. van kampen , stochastic processes in physics and chemistry . ( north - holland , amsterdam , 1992 ) h. mori , h. fujisaka , and h. shigematsu , prog .
51 , 109 ( 1974 ) .
b . zhu , phys .
a 42 , 3374 ( 1990 ) .
zhu , s. singh , j. lee , and g. w. robinson , chem .
152 , 221 ( 1991 ) .
m. f. gelin and d.s .
126 , 514721 ( 2007 ) .
s. ramakrishna and t. seideman , j. chem .
124 , 034101 ( 2006 ) .
the rotational fokker - planck equation with gaussian friction gives very similar results ( m. f. gelin and d. s. kosov , unpublished ) .
strictly speaking , @xmath116 has uncertainty of the kind @xmath283 if we put @xmath284 in eq .
( [ sne ] ) . if , however , we consider the limit of @xmath116 and @xmath124 when @xmath285 , then the quantities are well behaved . a. g. st .
pierre and w. a. steele , phys .
184 , 172 ( 1969 ) .
( full line ) and rotational energy relaxation time @xmath145 ( dashed line ) vs. parameter @xmath71 calculated in the generalized j - diffusion model ( @xmath139 ) for @xmath80 and @xmath82 .
the dotted line corresponds to the situation when the reduced value of the relaxation times equals @xmath133 . ]
( full lines ) and rotational energy cf @xmath153 ( dashed line ) calculated for the `` realistic values '' of the model parameters ( @xmath148 , @xmath80 , @xmath149 ) and @xmath150 .
the lower line corresponds to the standard j - diffusion model ( @xmath40 ) and the upper lines correspond to the generalized j - diffusion model ( @xmath139 ) . ]
k into the heat bath of argon atoms at @xmath166 k ( refs .
@xcite ) , and dotted lines show the results of the calculations within the generalized j - diffusion model ( @xmath128 , @xmath127 ) with the initial @xmath290-distribution ( [ deltad ] ) centered at @xmath168 . for cn at @xmath277
k , @xmath9@xmath278 ps .
, title="fig : " ] k into the heat bath of argon atoms at @xmath166 k ( refs .
@xcite ) , and dotted lines show the results of the calculations within the generalized j - diffusion model ( @xmath128 , @xmath127 ) with the initial @xmath290-distribution ( [ deltad ] ) centered at @xmath168 . for cn at @xmath277
k , @xmath9@xmath278 ps . , title="fig : " ] of the first ( a ) and second ( b ) rank vs. the collision rate @xmath24 calculated for the `` realistic values '' of the model parameters ( @xmath148 , @xmath80 , @xmath149 ) .
the dashed lines correspond to the @xmath179 and modernized hubbard asymptote @xmath291 $ ] calculated within the standard j - diffusion model ( @xmath40 ) .
the full lines correspond to their generalized j - diffusion model ( @xmath139 ) counterparts .
the standard hubbard asymptotes , @xmath292,$ ] are depicted by dotted lines .
, title="fig : " ] of the first ( a ) and second ( b ) rank vs. the collision rate @xmath24 calculated for the `` realistic values '' of the model parameters ( @xmath148 , @xmath80 , @xmath149 ) .
the dashed lines correspond to the @xmath179 and modernized hubbard asymptote @xmath291 $ ] calculated within the standard j - diffusion model ( @xmath40 ) .
the full lines correspond to their generalized j - diffusion model ( @xmath139 ) counterparts .
the standard hubbard asymptotes , @xmath292,$ ] are depicted by dotted lines .
, title="fig : " ] | in order to adequately describe molecular rotation far from equilibrium , we have generalized the j - diffusion model by allowing the rotational relaxation rate to be angular momentum dependent .
the calculated nonequilibrium rotational correlation functions ( cfs ) are shown to decay much slower than their equilibrium counterparts , and orientational cfs of hot molecules exhibit coherent behavior , which persists for several rotational periods . as distinct from the results of standard theories , rotational and orientational cfs
are found to dependent strongly on the nonequilibrium preparation of the molecular ensemble .
we predict the arrhenius energy dependence of rotational relaxation times and violation of the hubbard relations for orientational relaxation times .
the standard and generalized j - diffusion models are shown to be almost indistinguishable under equilibrium conditions . far from equilibrium ,
their predictions may differ dramatically . |
the origin of fast radio bursts ( frbs ) remains a mystery after one decade of observations @xcite .
the distance scale of frbs is finally settled to the cosmological range ( @xmath0 gpc ) thanks to multi - wavelength follow - up observations that nail down the host galaxy of the repeating source that produced frb 121102 @xcite .
the collected data for frbs so far are however perplexing and inconsistent . 1 . whereas the source of frb 121102 clearly repeats @xcite , other frbs
so far have not been observed to repeat , despite dedicated follow - up observations for some of them @xcite ; 2 . in the error circle of frb 150418
, a bright fading radio source was discovered in a galaxy at redshift @xmath1 , which was claimed as the afterglow of the frb @xcite .
however , the radio source re - brightened later , leading to the suggestion that the radio source is a coincident background agn not associated with the frb @xcite .
however , the chance probability to have such a highly variable agn in the field of view of the frb and to have a bright flare follow the frb is low @xcite ; 3 .
@xcite discovered a sub - threshold long grb that was spatially and temporally associated with frb 131104 @xcite ; 4 .
the repeating bursts from the frb 121102 source are not associated with bright transient counterparts in other wavelengths , but the repeater is rather associated with a steady bright radio * source * @xcite .
many frb models have been proposed so far .
however , no model can explain all the above observational facts . the catastrophic models ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) can not interpret the repeating frb 121102 .
the previously favored super - giant - pulse model ( in analogy to nanoshots from the crab pulsar ) @xcite invokes nearby galaxy not at the cosmological distances and therefore are challenged by the fact that the repeater is located in a host galaxy at redshift @xmath2 @xcite ( but see @xcite ) .
the magnetar giant flare model @xcite is in apparent conflict with the non - detection of bright radio pulse from the sgr 1806 - 20 giant flare @xcite .
other repeating models ( e.g. * ? ? ?
* ; * ? ? ?
* ) invoke specific physical conditions .
these models also can not account for the apparent association of frb 131104 with the gamma - ray transient swift j0644.5 - 5111 @xcite .
models invoking more energetic magnetars as the source of frbs have been proposed @xcite . however , in order to account for both the repeater and the putative association of a grb for frb 131104 @xcite , both catastrophic and repeating mechanisms are needed . here
we propose an alternative mechanism to produce frbs , which may give a unified interpretation of all frbs observed so far .
this model invokes an interaction between an astrophysical plasma stream and a foreground regular pulsar .
the pulsar is otherwise non - detectable at the cosmological distance , but may produce an frb as observed by an earth observer when the plasma stream with significant ram pressure combs the magnetosphere towards the anti - stream direction and sweeps the direction of earth .
the condition for a cosmic comb is that the ram pressure of the stream is greater than the magnetic pressure of the pulsar magnetosphere . for a most general expression
, we write the ram pressure in the relativistic form .
the energy - momentum tensor of a perfect fluid is given by @xmath3 , where @xmath4 is the metric tensor , @xmath5 is the dimensionless 4-velocity , @xmath6 is the dimensionless 3-velocity vector with amplitude @xmath7 , @xmath8 is the lorentz factor , @xmath9 , @xmath10 , and @xmath11 are the mass density , internal energy density , and pressure of the fluid , respectively , @xmath12 is the adiabatic index , and @xmath13 is speed of light . for a fluid with co - moving mass density @xmath14 , internal energy density @xmath10 and pressure @xmath15 ,
the lab - frame energy density is @xmath16 . subtracting the rest mass energy density one
can obtain the ram pressure .
the condition for a cosmic comb then reads @xmath17 for a non - relativistic ( @xmath18 ) , cold ( @xmath19 ) flow , this is reduced to the familiar form of @xmath20 the magnetosphere of a pulsar is enclosed within the light cylinder with radius @xmath21 , where @xmath22 and @xmath23 are the angular frequency and period of the pulsar , respectively .
the magnetosphere of a pulsar is significantly modified if the ram pressure exceeds the magnetic pressure at the light cylinder , @xmath24 , where @xmath25 cm is the radius of the neutron star , and a dipolar magnetic configuration has been assumed .
the right hand side of eqs.([eq : cond ] ) and ( [ eq : cond - nr ] ) is therefore @xmath26 where the median values of surface magnetic field @xmath27 and period @xmath28 s for galactic radio pulsars have been adopted ( throughout the paper , the convention @xmath29 is adopted in cgs units ) . noticing the strong dependence on @xmath30 and especially on @xmath23
, one can conclude that cosmic combs more easily happen in slow and low - field pulsars . for a relativistic blastwave such as a grb afterglow or a relativistic agn flare , since @xmath31 and @xmath32 due to relativistic shock heating , the left hand side of eq.([eq : cond ] ) can be simplified as @xmath33 for a nominal value @xmath34 and @xmath35 , where @xmath36 . for a non - relativistic outflow such as a supernova explosion , a quasi - isotropic compact star merger ejecta , or a quasi - isotropic tidal disruption ejecta , the left hand side of eq.([eq : cond - nr ] ) can be expressed as @xmath37 for an impulsive explosion ( where @xmath38 is the ejecta mass , @xmath39 is the thickness of the shell , @xmath40 is the distance from the center of explosion , and @xmath41 is the speed of the ejecta ) , or @xmath42 for a continuous wind ( where @xmath43 is the wind mass loss rate ) .
one can see that cosmic comb condition can be satisfied for a variety of systems if the pulsar is close enough to the source . for typical parameters ,
the horizon " of a cosmic comb is of the order of 0.1 pc .
however , noticing the sensitive dependence of magnetic pressure on @xmath23 and @xmath44 , the horizon can easily reach @xmath0 pc for slower and weaker - field pulsars .
if the cosmic comb condition is satisfied , when the plasma stream suddenly arrives , the pulsar magnetosphere is combed towards the anti - stream direction in a duration of @xmath45 , where @xmath46 is the distance of the sheath from the pulsar ( fig.1 , in analogy with the earth magnetosphere ) .
the strong ram pressure of the stream may trigger magnetic reconnections , which would quickly accelerate particles to relativistic speeds within a time scale much shorter than the combing time scale .
the relativistic particles move along the magnetic field lines and flow out from the sheath to produce coherent radio emission . during the combing process
, the sheath very quickly sweeps a near-@xmath47 hemisphere solid angle .
if the electrons ( or electron - positron pairs ) in the sheath move with a lorentz factor @xmath48 , then the emission beam angle of the sheath would be @xmath49 , which is usually much smaller than the physical opening angle of the sheath . if the distance range over which emission into the characteristic radio band of the telescope ( e.g. ghz range ) is small , an observer at a random direction in the `` night '' side of the hemisphere would detect the signal in a duration @xmath50 this is consistent with the millisecond durations of the observed frbs if @xmath51 is satisfied .
the coherent radio emission mechanism is not specified .
one interesting estimate is that with the nominal parameters , the curvature radiation frequency is @xmath52 which falls into the range of frb detection frequency ( the curvature radius @xmath53 has been assumed ) . if the combing process may allow particles to emit in phase ( the antenna mechanism ) , bright coherent radio emission can be generated .
alternatively , a cyclotron instability mechanism , as was proposed to interpret pulsar radio emission @xcite , may operate near the light cylinder region to power the frb .
a sheath - origin coherent emission has been introduced to interpret radio emission of pulsar b in the double pulsar system @xcite . here
we assume that the combing process can produce bright coherent radio emission detectable from cosmological distances , and encourage a full investigation of the coherent mechanism of such a process in detail .
this model can explain the puzzling properties of several frbs in a unified way : * frb 150418 : * @xcite discovered a bright radio source starting from 2 hours after the frb .
the source dimmed in the next two observational epochs at 6 and 8 days after the frb , respectively , and reached a quiescent level with a steady flux .
@xcite discovered that the source re - brightened to the level of the original detection flux more than 300 days later , suggesting that the radio source is an agn rather than the afterglow of the frb .
long term monitoring of the source @xcite suggested that the source is usually not in the high state .
a monte carlo simulation suggests that the random probability of having the frb to occur almost during the peak flux time of the agn activity is very low , i.e. @xmath54 @xcite . instead of attributing the agn flare to an independent event from the frb , we interpret frb 150418 as emission from a combed pulsar by the agn flare .
a prediction is that frb 150418 _ may _ repeat during another bright flare from the same agn .
however , not all flares may trigger additional frbs from the same pulsar .
this is because at the close distance ( @xmath55 pc ) from the super - massive black hole , the pulsar must be undergoing orbital motion , so that there are occasions when the geometry does not work for the cosmic comb signal to be detectable from earth . within this picture ,
the galaxy at @xmath1 is indeed associated with frb 150418 , as is supported by the measured dm of the frb @xcite . *
frb 131104 : * @xcite discovered a sub - threshold , putative grb that coincides with frb 131104 both in spatial position and in time .
a radio afterglow was not detected @xcite , but the non - detection is consistent with the afterglow model if the ambient density is low ( as expected from the ns - ns or ns - bh merger models ) or the shock microphysics parameters are low @xcite . the possible mechanisms to produce an frb associated with a grb include collapse of a supra - massive millisecond magnetar to a black hole @xcite , which requires that the frb appears near the end of an extended x - ray plateau ; or a pre - merger electromagnetic processes @xcite , which requires that the frb leads the burst .
the latter scenario may be argued to marginally match the data @xcite .
however , there might be @xmath56-ray emission already 7 seconds before the frb according to the data .
furthermore , the swift bat was not pointing toward the source direction before @xmath57 seconds with respect to the frb @xcite .
so it is likely that the frb occurred during the process of a long - duration grb .
if so , known models are difficult to interpret the frb . in the cosmic comb model ,
one requires that a pulsar is located at a distance @xmath58 away from the central engine in the direction of the jet ( or at a closer distance if the line of sight is mis - aligned from the pulsar - engine direction ) . considering a possible star forming region for a long grb or a possible globular cluster for a ns - ns or ns - bh merger event ,
the chance probability to have a foreground pulsar from the grb may not be small .
* the repeater ( frb 121102 ) : * the repeater is located in a star - forming dwarf galaxy at @xmath59 @xcite .
the source is associated with a radio source @xcite , which is offset from the center of the galaxy @xcite .
a plausible scenario might be that the source of the frb , likely a rapidly spinning magnetar , is at the center of the radio source and pumping energy to power a nebula ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. however , this model predicts an observable evolution of dm over the year time scale @xcite , which is marginally inconsistent with the non - detection of dm evolution of the repeating frbs . within the cosmic comb scenario proposed in this paper ,
the repeating bursts may originate from a foreground pulsar being episodically combed by an unsteady flow from a young supernova remnant .
if the condition ( [ eq : cond ] ) is marginally satisfied , the pulsar may relax to its normal magnetospheric configuration after a particular combing , but may be combed again and again when clumps with a higher ram pressure reach the pulsar magnetosphere region repeatedly .
the pulsar is therefore observed to emit frbs repeatedly . for a remnant with a finite width @xmath39 and speed @xmath41 ,
the repeating phase may last for @xmath60 .
since the repeater has been observed to repeat in a multi - year time scale , the remnant may be continuously energized by a central engine , likely a rapidly rotating neutron star .
since the frb source is a foreground pulsar from the central source , the dm evolution could be much weaker depending on the geometry , consistent with the data .
there is no direct observational evidence of ram pressure variation within a supernova remnant .
however , for a nebular powered by continuous energy injection from a central engine ( which is _ not _ the source of frbs in the cosmic comb model ) , variation of ram pressure of the stream is expected . for a marginally satisfied comb condition envisaged here ,
a variation of ram pressure by a factor of a few would suffice to make a repeating frb source as observed .
* other frbs : * no counterparts have been claimed for other frbs . within the cosmic comb model ,
the ram pressure of the plasma stream essentially depends on the energy _
flux _ of the stream source .
for example , a flare from a companion star ( similar to a corona mass ejection event of the sun ) may provide a comparable ram pressure to a pulsar as the blastwave of a more distant grb or supernova .
as a result , one does not necessarily expect that all frbs are associated with bright counterparts .
it is possible that a fraction of frbs may be associated with grbs , supernovae , agn flares , or tidal disruption events , while other frbs may be combed by a stellar flare of a companion star so that no bright counterpart is detectable .
we have proposed a scenario of frbs that may interpret the perplexing , inconsistent frb phenomenology within a unified framework .
the magnetosphere of a slow , low - field pulsar may be combed by a cosmic plasma stream towards the direction of the earth to make an frb .
the stream may originate from an explosion ( a supernova , a grb , or a neutron star merger event ) , an agn flare , a tidal disruption event , or even a stellar flare from a binary companion of the pulsar .
the event rate density of frbs is estimated as @xmath61 , where @xmath62 is the number of frbs per day all sky with the current telescope sensitivity , and the typical value @xmath63 is scaled to @xmath64 @xcite .
the cosmic comb model can easily account for such a large event rate density .
this is because many explosions ( e.g. supernovae and grbs ) and flares ( from agns or companion stars ) can potentially make frbs if there are one or more pulsars in the vicinity of each plasma stream .
in fact , the total event rate density of all the relevant explosions and flares exceed by orders of magnitude than the inferred frb event rate density .
a correct frb event rate density may be obtained if the probability of having a slow , weak field pulsar in the vicinity is not negligibly small . in star forming regions ( where grbs and supernovae occur ) , globular clusters ( where compact star mergers occur ) , and galactic center regions ( where agn flares occur ) , the stellar density can be as high as 100 - 1000 @xmath65 @xcite .
since one old pulsar within @xmath0(0.1-several ) pc from an explosion source or agn would give a 50% probability to make a cosmic comb towards earth ( as long as earth is in the night side hemisphere of the stream ) , one would expect a good fraction of explosions / agn flares to trigger cosmic combs .
adding binary systems composed of an old pulsar and a flaring companion , the event rate would be even higher . 1 per century based on the observed frb rate .
] more detailed monte carlo simulations are needed to better estimate the event rate density of cosmic combs
. supports to the cosmic comb model proposed in this paper may be obtained by a list of observational tests , e.g. a detection of repeating bursts from frb 150418 and the localization of the bursting source to the spatially coincident agn ; more detections of frbs associated with grbs and afterglows ; detections of frbs that are associated with supernovae , tidal disruption events , and other cosmological transients . on the other hand , it is entirely possible that multiple mechanisms are at play to make frbs .
some other models have specific predictions that can be tested by future observations .
for example , if future localizations of frbs all reveal associations with bright radio nebulae , it would suggest a new - born millisecond magnetar as the source of frbs ( e.g. * ? ? ? * ; * ? ? ? * ) , an frb coincident with a gravitational wave ( gw ) signal due to a neutron star - neutron star ( ns - ns ) , neutron star - black hole ( ns - bh ) , or black hole - black hole ( bh - bh ) merger would suggest a direct connection between the frb and the merger @xcite ; and an frb coincident with the disappearance of an extended plateau emission would suggest the ejection of a magnetosphere during the collapse of a supra - massive neutron star ( `` blitzar '' ) as the mechanism of frbs @xcite . | recent observations of fast radio bursts ( frbs ) indicate a perplexing , inconsistent picture . we propose a unified scenario to interpret diverse frbs observed .
a regular pulsar , otherwise unnoticeable at a cosmological distance , may produce a bright frb if its magnetosphere is suddenly `` combed '' by a nearby , strong plasma stream towards the anti - stream direction .
if the earth is to the night side of the stream , the combed magnetic sheath would sweep across the direction of earth and make a detectable frb .
the stream could be an agn flare , a grb or supernova blastwave , a tidal disruption event , or even a stellar flare . since it is the energy _ flux _ received by the pulsar rather than the luminosity of the stream origin that defines the properties of the frb , this model predicts a variety of counterparts of frbs , including a possible connection between frb 150418 and an agn flare , a possible connection between frb 131104 and a weak grb , a steady radio nebula associated with the repeating frb 121102 , and probably no bright counterparts for some frbs . |
microscopic traps provide a powerful tool for the control and manipulation of bose - einstein condensates over micrometer distances .
microstructured surfaces , known as atom chips , are particularly interesting for this purpose since they can be tailored to provide a variety of trapping geometries @xcite and promise well - controlled quantum state manipulations of neutral atoms in integrated and scalable microtrap arrays .
ultimately there is the possibility of controlling the quantum coherences within arrays of individual atoms for use in quantum information processing @xcite .
this technology is attractive because it appears robust and scalable and because trapped neutral atoms can have long coherence times .
however , atoms in these traps are held close to the micro - structured material surfaces , which are typically at room temperature .
thermal fluctuations give rise to johnson noise currents in the material @xcite .
such currents are normally observed as a noise voltage across a resistor , but they also cause the electromagnetic field near a conducting solid to fluctuate with a broad noise spectrum . for atoms trapped close to the surface of a conductor
these fluctuating fields can be strong enough to drive rf magnetic dipole transitions that flip the atomic spin .
if the atom is in a magnetic trap where only low - field - seeking zeeman sublevels are confined , the spin flips lead to atom loss .
this is known experimentally @xcite as well as theoretically @xcite .
the loss rate increases strongly as the atoms approach the metallic surface of an atom chip . for a given desired lifetime , this restricts how close the trapped atoms can be brought to the surface , which in turn determines the period of the smallest trapping structures that can be imposed on the atom by the chip .
the paper is organized as follows : in section [ bas_eq_q ] we introduce the basic equations and discuss the quantization of an electromagnetic field in the presence of a dispersing and absorbing dielectric body
. then , in section [ gamma_sec ] , we derive a general expression for the spontaneous and thermal spin - flip rates of an atom due to the coupling of its magnetic moment to the magnetic field . this derivation is based on the zeeman hamiltonian of the system and the corresponding heisenberg equations of motion .
we show that the spin - flip rate is determined by the dyadic green tensor of the classical , phenomenological maxwell equations . in section [ g_sec ]
we present the scattering green tensor for a @xmath0-layer cylindrical body surrounded by an unbounded homogeneous medium , with details being given in appendix [ r_appendix ] .
then , in section [ final formula ] , we use this green tensor to obtain an explicit analytical expression for the total spin - flip rate of an atom above a @xmath0-layer wire .
some numerical results are presented and discussed in section [ num_sec ] .
the numerical results are compared with the corresponding results for a slab and with the experimental measurements presented by jones et al . in ref .
our conclusions are given in section [ conclusions ] .
in classical electrodynamics , dielectric matter is commonly described in terms of a phenomenologically introduced dielectric susceptibility .
let us consider a classical electromagnetic field , described by the phenomenological maxwell s equations , without external sources .
we restrict our attention to isotropic but arbitrarily inhomogeneous non - magnetic media , and assume that the polarization responds linearly and locally to the electric field . a linear response formalism similar to that presented below
can also be found in refs .
@xcite . the most general relation between the matter polarization and the electric field consistent with causality and the fluctuation - dissipation theorem is @xcite [ p_rt ] * p*(*r*,t ) = _ 0 _ 0^ d ( * r * , ) * e*(*r*,t- ) + * p*_n(*r*,t ) , where @xmath1 is the linear susceptibility .
the inclusion of the noise polarization @xmath2 is necessary to fulfil the fluctuation - dissipation theorem .
it is this fluctuating part of the polarization that is unavoidably connected with the loss in the medium . converting the displacement field @xmath3 into fourier space using , we obtain ( * r * , ) = _ 0 ( * r * , ) * e*(*r * , ) + * p*_n(*r * , ) , where @xmath4 is the complex permittivity and @xmath5 is the temporal fourier transform of @xmath1 .
the real part of the permittivity ( @xmath6 , responsible for dispersion ) and the imaginary part ( @xmath7 , responsible for absorption ) are related to each other by the kramers - kronig relation . using maxwell s equations in fourier space , we find that @xmath8 satisfies the helmholtz equation @xmath9 with the solution ( * r * , ) = ^2 _ 0 d^3 * r* ( * r*,*r* , ) _ n(*r* , ) , where the green tensor @xmath10 is a second - rank tensor determined by the partial differential equation @xmath11 where @xmath12 is the unit dyad . together with the boundary condition at infinity
, this equation has a unique solution . in accordance with maxwell s equations
the corresponding solution for the magnetic field in fourier space is @xmath13 .
as we have seen , the noise polarization @xmath2 plays a fundamental rle in determining the electric field .
the form of @xmath2 follows from the fluctuation - dissipation theorem , which states that the fluctuations of the macroscopic polarization are given by the imaginary part of the response function [ here @xmath14 .
if we pull out a factor and define the dynamical variables @xmath15 as the fundamental @xmath16-correlated random process , we find that we can write the noise polarization as @xcite [ p_n ] * p*_n(*r * , ) = i ( * r * , ) . upon quantization
, we replace the classical fields @xmath17 by the operator - valued bosonic fields @xmath18 which we associate with the elementary excitations of the system composed of the electromagnetic field and the absorbing dielectric matter .
they satisfy the equal - time commutation relations @xmath19 $ ] @xmath20
. the magnetic - field operator in the schrdinger picture can now be obtained as ( * r * ) = ^(+)(*r * ) + ^(-)(*r * ) , ^(-)(*r * ) = [ ^(+)(*r*)]^where ^(+)(*r * ) = _ 0^ d ( * r * , ) , is its positive - frequency part . in this way , the electromagnetic field is expressed in terms of the classical green tensor satisfying the helmholtz equation ( [ g_helm ] ) and the continuum of the fundamental bosonic field variables @xmath21 .
all the information about the dielectric matter is contained , via the permittivity @xmath22 , in the green tensor of the classical problem .
we close this section by mentioning two important properties of the green tensor . it can be shown that the ( onsager ) reciprocity relation @xmath23 holds @xcite .
additionally , another useful property is the integral relation [ g_rel ] d^3 * r* _ i(*r* , ) g_kl(*r*,*r* , ) g_nl^(*r*_a,*r* , ) + = g_kn(*r*,*r*_a , ) , & & which we will use later in this paper .
both relations essentially follow from linear response theory , with being equivalent to the fluctuation - dissipation theorem @xcite .
it should be noted that we assume the dielectric permittivity to possess at least an infinitesimal imaginary part everywhere to avoid surface contributions in eq .
( [ g_rel ] ) .
the hamiltonian of the combined system of electromagnetic field and absorbing matter , from which the ( quantized ) phenomenological maxwell s equations can be derived , can be written in terms of the basic field operators @xmath18 in the diagonal form @xmath24 which leads in the heisenberg picture to the ( quasi - free ) time evolution @xmath18 @xmath25 @xmath26 .
here we have also included an atom through the operators @xmath27 and the energy @xmath28 of the atomic state @xmath29 ( @xmath30 ) .
the interaction of the atom at position @xmath31 with a magnetic field @xmath32 is described by the zeeman hamiltonian @xmath33 , where @xmath34 is the magnetic moment operator associated with the transition @xmath35 .
the magnetic moment vector is = i | _ b ( g_s + g_l - g_i ) | f , where @xmath36 is the bohr magneton , @xmath37 is the electronic spin operator , @xmath38 is the orbital angular momentum operator , @xmath39 is the nuclear spin operator and @xmath40 , @xmath41 and @xmath42 are the corresponding @xmath43-factors . we restrict our attention to @xmath44 , which corresponds to the ground state of an alkali atom , and we neglect the small nuclear magnetic moment in comparison with the bohr magneton . in the rotating - wave approximation , we can then write the zeeman hamiltonian as @xmath45 \,,\ ] ] where the atomic raising ( lowering ) operator @xmath46 [ @xmath47 satisfies the commutation relation @xmath48=\pm \hat{\xi}^{(\pm)}$ ] , with @xmath49 @xmath50 . repeated indices @xmath51 indicate a sum over spatial vector components . using the hamiltonian
( [ h_rwa ] ) , the heisenberg equation of motion for the atomic quantity @xmath52 is given by [ mu_dl ] _
z(t ) = - f|_q|i ^(+ ) _
q^(+)(*r*_a ) + . furthermore ,
the heisenberg equation of motion for the bosonic field operator is [ f_dl ] ( * r*,,t)&= & - i_i(*r * , , t ) + & & + i|_q|f ^(- ) _ qpj _ p g_ji^(*r*_a,*r * , ) , where @xmath53 is the levi - civita symbol and @xmath54 .
this equation can now be formally integrated to yield [ f_markov ] + & & i|_q|f _ qpj _ p g_ji^(*r*_a,*r * , ) , where @xmath55 denotes the freely evolving basic - field operators .
the lowering operator @xmath56 in can be found by solving its heisenberg equation of motion . in the markov approximation
, this solution can be reduced to its slowly varying part @xmath57 in so that the time integral can be approximated by & & _
i(*r*,,t ) = _ i,(*r*,,t ) + i|_q|f ^(-)(t ) + & & _ qpj _ p g_ji^(*r*_a,*r * , ) ( _ if- ) , where @xmath58 ( @xmath59 denotes the principal value ) and @xmath60 is the transition frequency corresponding to the flip @xmath61 in the atom s internal state . substituting this formal solution into the expression for the magnetic field , we obtain [ b_sol ] + & & + i|_p|f_qjk _ pmn _ j _ m ^(-)(t ) ( _ if- ) + & & d^3*r * _ i(*r * , ) g_kl(*r*_a,*r * , )
g_nl^(*r*_a,*r * , ) .
the spatial integral can be evaluated using the integral relation yielding @xmath62 .
therefore , becomes + & & + i|_k|f ^(-)(t ) ( _ if- ) + & & [ ( * r*_a,*r*_a , ) ] _ qk .
performing the @xmath63-integration and inserting into , we obtain + & & + , where the spontaneous spin - flip rate @xmath64 arises from the @xmath16 function ( the real part of the @xmath65 function ) and is given by [ gamma_b_generall ] + & & [ ( * r*_a , * r*_a , _ if ) ] _ qp , and where the term @xmath66 arises from the principal - value integral ( the imaginary part of the @xmath65 function ) and is identified as the radiative frequency shift .
furthermore , the shifted frequency is given by @xmath67 .
in what follows , the transition frequency is always taken to be the shifted frequency @xmath68 that one measures in an experiment and not the bare frequency @xmath69 .
for simplicity , we omit the tilde in all subsequent formulas .
note that the same result for @xmath70 is obtained when using an appropriately derived master equation as done in ref .
@xcite .
we assume that the dielectric body is in thermal equilibrium with its surroundings .
the magnetic field is then in a thermal state with a temperature @xmath71 , equal to the temperature of the dielectric body . the total flip rate for the atom is therefore given by @xmath72 , where the mean thermal occupation number is [ n_th ] _
= , and @xmath73 is boltzmann s constant . at zero temperature ,
i.e. @xmath74 , the relaxation dynamics is entirely due to the spontaneous flip rate @xmath70 . for large @xmath71 on the other hand ,
@xmath75 @xmath76 and the spin flip rate is predominantly induced by thermal fluctuations . in the experiment of ref .
@xcite @xmath77rb atoms are initially pumped into the trapped state @xmath78 .
thermal fluctuations of the magnetic field then cause the atoms to evolve into hyperfine sublevels with lower @xmath79 . upon making a transition to the @xmath80 state , the atoms are more weakly trapped and are largely lost from the region of observation , causing the measured atom number to decay with a rate @xmath81 .
here we are introducing the notation @xmath82 for the total spin - flip rate associated with the transition @xmath83 .
the geometry we are considering in this paper is a @xmath0-layer cylinder surrounded by an unbounded homogeneous medium ( see ) .
this corresponds to the experimental geometry in ref .
@xcite . -layer cylinder surrounded by an unbounded homogeneous medium .
the outer region is labelled layer @xmath84 ( vacuum ) , the coating is layer @xmath0 and the cylinder core is layer @xmath85 .
the distance from the surface of the outermost layer to the atom is @xmath86.,width=170 ] ( 0,0 ) ( -4,64)@xmath87 ( -10,63)@xmath88 ( 14.8,57.6)layer @xmath85 ( 8,72)layer @xmath0 ( -35,70)layer @xmath84 ( -36,67)(vacuum ) ( -25,25)@xmath89 ( 16,35)@xmath90 ( -6,51)@xmath91 ( 13,51)@xmath86 ( 18,43)atom because the helmholtz equation is linear , the associated green tensor can be written as a sum , [ g_tot ] ( * r*,*r*,)= ^0(*r*,*r* , ) + ^(*r*,*r* , ) , where @xmath92 represents the contribution from the vacuum and @xmath93 describes the part due to the wire . when the atom is located in layer 3 ,
the scattering contribution is @xcite @xmath94 where [ cal_r_n ] + & & r_n^11(h ) + & + & r^12_n(h ) ( - ) + & & + & + & r^22_n(h ) . for simplicity
, we have omitted the tensor product symbol @xmath95 between the _ e_ven and _ o_dd cylindrical vector functions defined by @xmath96 $ ] and @xmath97/k_3 $ ] . the scalar eigenfunctions @xmath98 satisfy the homogeneous scalar wave equation @xcite .
it follows from these definitions that [ m_def ] _
o n(h ) & = & k_3 * n*_^e _ o n(h ) , + [ n_def ] _
o n(h ) & = & k_3 * m*_^e _ o n(h ) .
explicitly , [ n_vec ] + & & ih z_n(_3 ) + ( n ) * e * _ + _ 3 ^ 2 z_n(_3 ) + ( n ) * e*_z ] e^i h z , [ m_vec ] + & & - + ( n ) * e * _ ] e^i h z .
the primes in indicate the spherical coordinates @xmath99 .
the superscript @xmath100 indicates that @xmath101 should be replaced by the hankel function of first kind @xmath102 . otherwise , @xmath101 is the bessel function of first kind @xmath103 .
the propagation constant in the @xmath104 direction is @xmath105 , where @xmath106 is the wave number of the @xmath107th layer .
the permittivity of the @xmath107th layer is denoted by @xmath108 .
the scattering reflection coefficients @xmath109 are given in appendix [ r_appendix ] ( @xmath110 ) .
the double curl of the green tensor in can be written [ curl_curl_g ] & & ^(*r*,*r* , ) = + & & _ n=0^ ( 2-_0n ) [ 2_curl ] , where [ 2_curl_g_sum ] * i * _ n _ n ( * r*,*r* , ) = _ -^ d h _ n(h ) .
note that the curls are computed by replacing @xmath111 by @xmath112 and vice versa , according to eqs .
( [ m_def ] ) and ( [ n_def ] ) .
also note that the integration variable @xmath113 is the wave number in the @xmath91-direction ( see ) . from the symmetry of the integrand , it is easy to show that @xmath114 , where @xmath115 @xmath116 ( @xmath117 ) .
note that the ( onsager ) reciprocity relation as mentioned in section [ bas_eq_q ] implies that @xmath118 .
the spin - flip rate in free space is readily derived from since @xmath119 @xmath120_{qp}$ ] @xmath121 , where @xmath122 is the free space wave number corresponding to the atomic transition .
we use the notation @xmath123 here because in our discussion of the cylindrical wire , the third layer is a vacuum .
hence [ gamma_0 ] ^ 0_if = _ 0 k_3 ^ 3 s_if^ 2 , where we have introduced the angular factor @xmath124 @xmath125 @xmath126 since we are interested here in a spin flip , which by definition changes @xmath79 .
we have moreover used the fact that the two transverse matrix elements are equal in absolute value as a result of symmetry . for the @xmath77rb ground state transition @xmath127 .. ] in order to find the contribution of the wire to the spontaneous spin - flip rate , we use eqs .
( [ gamma_b_generall ] ) and ( [ curl_curl_g ] ) .
the quantization axis is taken to be along the @xmath91 direction , corresponding to the direction of the bias field that the trapped atoms experience in the experiment .
we obtain [ gamma_3_dir ] ^_if= ^ 0_if _ n=0^ ( 2-_0n ) . here
we have once again used the facts that @xmath128 .
the dimensionless integrals @xmath129 are given by & & [ i_xx_chew ] ( _ n^)_xx + ( _ n^)_yy = _ -^ d q + & & \ { [ r_n^11(q ) + q^2 r_n^22(q ) ] [ ( h_n3)^2 + ( _ 3 h_n3)^2 ] + & & + 2 i q r_n^12(q ) ( - ) _ 3 ( h_n3 ^ 2) } , since @xmath130 .
we have used the simplified notation @xmath131 and the primes in denote the derivative with respect to the full argument of the relevant function , e.g. @xmath132 .
we have also chosen to write the permittivity of the @xmath107th layer relative to the outermost layer , i.e. @xmath133 . the wave number for layer @xmath107 is then given by @xmath134 , and the dimensionless propagation constant @xmath135 in the @xmath104 direction can be written as @xmath136 , where @xmath137 is the dimensionless integration variable .
the skin depth is the characteristic length scale on which an electromagnetic wave is damped within a conducting medium .
it is given by @xmath138 ( see e.g. ref .
@xcite ) , where @xmath139 is the resistivity of layer @xmath107 .
since the spin - flip frequency is very much lower than the resonance frequencies of the material in the wire , the relative permittivity is related to the skin depth by @xcite [ eps_1_approx ] _
p^ = i .
we see from that the total spin - flip rate is equal to the sum of the free space contribution and the scattering contribution . the total spin - flip rate for the rate - limiting transition @xmath140 is therefore [ gamma_tot ] ^b_21 = ( ^ 0 _ 21 + ^ _ 21 ) ( _ + 1 ) .
in the experiment of ref.@xcite , cold atoms are held in a microscopic trap near a current - carrying wire assumed to be at room temperature . the lifetime for atoms to remain in the microtrap
is measured over a range of distances down to @xmath141 m from the surface of the wire .
the wire consists of a central copper core with radius @xmath142 m and a @xmath143 m thick aluminium layer , i.e. @xmath144 m .
using , the resistivities @xmath145 m for cu and @xmath146 m for al give skin depths of @xmath147 m for cu and @xmath148 m for al at frequency @xmath149 khz .
( 0,0 ) ( -130,80)@xmath150 ( s ) ( -10,-6)@xmath86 ( @xmath151 m ) the dotted line in shows the lifetime
@xmath152 that we have calculated assuming a temperature of @xmath153k , together with the measured lifetimes ( crosses ) .
we see that the experimental results are close to the theory , indicating that the thermal spin flip mechanism is the primary cause of atom loss in the experiment .
nevertheless there is also a clear systematic discrepancy , with the measured lifetimes being @xmath154 shorter than expected .
we find excellent agreement when the temperature in our theory is increased to @xmath155k , as shown by the solid curve in .
we have re - examined the conditions under which the experiment was run and consider it most likely that the wire temperature was indeed @xmath156k , rather than the @xmath153k previously assumed .
such a temperature rise would be consistent with known power dissipation and with reasonable assumptions about the heat flow . in effect , the thermally driven spin flips have allowed us to measure the temperature of the wire ! the theory for the decay rate of an atom above a plane , thick slab is already known @xcite . applying this theory to an al slab with skin depth @xmath157 m and temperature @xmath155k
, we obtain the result shown dot - dashed in .
this curve lies below that for the wire , simply reflecting the fact that the slab contains a larger volume of fluctuating polarization than the wire . naturally , the two @xmath155k curves converge at sufficiently small atom - surface distances ( @xmath158 , @xmath159 , @xmath87 ) , and in that range they vary linearly with distance @xcite .
( 0,0 ) ( -130,80)@xmath150 ( s ) ( -10,-6)@xmath160 ( @xmath151 m ) the lifetime for the atom to remain in the trap exhibits a minimum with respect to variation of the skin depth , as illustrated in fig.[delta_2_i_fig ] , where the skin depth of the wire core @xmath159 is fixed at @xmath161 m but the skin depth of the outer layer @xmath160 is varied . below the minimum at @xmath162 m ,
a decrease of skin depth leads to an increase of lifetime in proportion to @xmath163 .
this happens despite a growth in the polarization noise [ see eqs .
( [ p_n ] ) and ( [ eps_1_approx ] ) ] because the region generating the noise is becoming thinner . in the small @xmath160
limit the outer layer approaches a perfect conductor , the core wire does not play any role , and the lifetime becomes exceedingly long .
by contrast , when the skin depth increases above @xmath164 m , the reduction in polarization noise is more influential than the growth of the source volume . in this region
it is the worse conductor that gives the longer lifetime . at large @xmath160 , the outer layer of the wire approaches the free space limit , and the lifetime is entirely determined by the skin depth and radius of the core . from a practical viewpoint it would normally be desirable to avoid the minimum of the lifetime curve
this means avoiding surface materials whose skin depth at the spin flip frequency is comparable with the atom - surface distance .
this is a generic result .
for example , at height z above a slab , one obtains the shortest lifetime when the skin depth is @xmath165 , as is readily derived from equation ( 23 ) of ref .
@xcite .
( 0,0 ) ( -117,80)@xmath150 ( -109,80)(s ) ( -2,-6)@xmath86 ( @xmath151 m ) from the same perspective of cold atoms trapped above small integrated circuits ( atom chips ) it is also interesting to see how the lifetime is altered when the dimensions @xmath88 , @xmath87 of the wire are varied or the atom - surface distance @xmath86 changes .
for example , let us scale all three lengths together , such that @xmath166 and @xmath167 , while the skin depths are fixed .
the result of such a scaling is illustrated in .
when the atom - surface distance is large compared with the skin depth , i.e. @xmath168 m , the spin - flip lifetime scales as @xmath169 .
this has the same exponent as the @xmath170 scaling of lifetime that applies at distance @xmath91 from a slab in the range where @xmath171 @xcite .
the correspondence seems natural to us since the wire is essentially a curved slab when the skin depth is small .
for a given ratio of atom - surface distance to wire size , the two lifetimes should therefore be related by a constant geometrical factor , resulting in the same distance scaling . at the opposite extreme , where @xmath172 , the slab result is @xmath173 .
by contrast , we see in that the lifetime outside the wire approaches a constant when @xmath174 .
this difference occurs because the thickness of the source region is not the skin depth , but rather the diameter of the wire , which we are scaling linearly with @xmath86 . in a similar way ,
it is possible to lengthen the lifetime of an atom above a slab by reducing the thickness of the slab to less than the skin depth @xcite . as a second example of scaling , we change the diameter of the wire , keeping @xmath175 but fixing the distance from the surface at @xmath176 m .
once again the skin depths are fixed .
the resulting variation in the lifetime of the atom with wire size is shown in . at large wire diameter , the lifetime approaches @xmath177 s , which is of course the same as the lifetime @xmath176 m above a slab with @xmath178 m skin depth .
by contrast , when the wire size is small , i.e. @xmath179 , the decreasing volume of material leads to a @xmath180 scaling of the lifetime . in the limit @xmath181 , @xmath182 vanishes and the lifetime for the atoms to remain in the trap is just the free - space blackbody rate given by the first term in . for @xmath183 khz and @xmath184 k ,
this free space lifetime is an astonishing @xmath185 s ( see also @xcite ) .
this figure emphasizes the very low strength of the electromagnetic field fluctuations in free space compared with those near a dielectric medium due to the surface modes .
( 0,0 ) ( -126,80)@xmath150 ( -117,80)(s ) ( -8,-6)@xmath87 ( @xmath151 m )
in this paper we have derived the magnetic spin flip rate for an atom close to an absorbing dielectric body . the rate is given in terms of a dyadic green tensor , allowing the expression to be applied in principle to a dielectric body of any shape .
we derive an explicit expression for the spin - flip rate of an atom outside a 2-layer cylindrical wire , as used in the experiment of jones _
we compare our numerical results with their measurements and we find lifetimes marginally longer than those measured in the experiment .
the most likely explanation for this discrepancy is that the wire was hotter than previously thought .
we also compare the cylindrical case with that of a slab and show that the spin - flip lifetime is systematically longer above a cylinder , as one would expect .
we have investigated how the lifetime of the atoms depends on the skin depth of the material .
we find the generic result that there is a minimum in the lifetime when the skin depth is comparable with the atom - surface distance .
when the dimensions of the wire and the atom - surface distance @xmath86 are varied together , the lifetime scales as @xmath186 at large @xmath86 , following the same scaling law as a corresponding plane , thick slab , whereas the lifetime approaches a constant at small @xmath86 .
if instead we fix the atom - surface distance and vary only the dimensions of the wire , the lifetime scales as @xmath187 when the wire is small , leaving only the very weak free - space decay rate in the limit of a vanishing wire diameter .
the main conclusion for atom chip design is that one should avoid a material whose skin depth at the spin flip transition frequency is comparable with the atom - surface distance
. the lifetime can also be improved by making sure that metal films on the surface are thinner than the skin depth .
we are indebted to m.p.a .
jones for valuable helpful comments .
this work was supported by the uk epsrc and by the fastnet and qgates networks of the eu .
p.k.r . acknowledges support by the research council of norway . s.s .
acknowledges support by the alexander von humboldt foundation .
the scattering reflection coefficients for a cylindrical geometry can be computed for any number of layers ( see e.g. refs .
@xcite ) . in this appendix
we present the explicit expressions for the scattering reflection coefficients corresponding to our @xmath84-layer cylindrical geometry . to find these reflection coefficients
we have used the iteration tensor equations in ref .
these iteration equations are given for arbitrary complex permittivity @xmath108 and arbitrary complex permeability @xmath188 .
therefore , the reflection coefficients presented in this appendix apply to arbitrary @xmath108 _ and _ arbitrary @xmath188 .
however , we stress that the theory presented in the main body of this paper is particular to non - magnetic media ; we assume that @xmath189 in all the layers @xmath107 .
the reflection coefficients are given as follows : [ r_11_final ] r_n^11(h ) & = & + & + & ( ) ^2 _ 3 ^ 2 _ 2 ^ 2 _ 3 , + [ r_12_final ] r_n^12(h ) & = & + & + & ( ) ^2 _ 3 ^ 2 _ 2 ^ 2 _ 3 , where t_n^11 & = & d_n32_n -t_n21 _ n , + t_n & = & d_n32 _ n - t_n21 _ n , and n_n & = & ( d_n32 ) ^2 .
moreover , we have [ alpha_a_def ] _ n & = & - ( a^ h_3 j_2 _ n _ 3 ) ^2 _ 2 b_11 + ( b^ h_3 j_2 _
n ) ^2_2 b_22 + & & - 2 _ 2 b_12 a^ h_3 j_2 _ n _ 3 b^ h_3 j_2 _ n , + [ beta_a_def ] _ n & = & - ( a^ h_3 j_2 _ n _ 3 ) ^2 _ 2 a_22 + ( b^ h_3 j_2 _
n ) ^2 _ 2 a_11 + & & + 2 _ 2 a_12 a^ h_3j_2 _ n_3 b^h_3 j_2_n , + [ gamma_def ] _ n&= & - a^ h_3 j_2 _ n _ 3 b^ h_3 j_2 _ n _ 3 b_11 - a^ h_3 j_2 _ n _ 3 b^ h_3 j_2 _ n _ 2 b_22 + & & + _ 2 b_12 [ a^ h_3j_2_n_3
a^ h_3 j_2 _ n_3 - ( b^ h_3 j_2_n ) ^2 ] , + [ delta_def ] _ n & = & - a^ h_3 j_2 _ n _ 3 b^h_3 j_2_n a_22 -a^ h_3 j_2 _ n _ 3 b^ h_3 j_2 _ n _ 2 a_11 + & & - _ 2 a_12 [ a^h_3 j_2_n_3 a^h_3j_2_n _ 3 - ( b^h_3 j_2_n ) ^2 ] , and [ a_11_def ] a_11 & = & a^ h_3 j_2 _ n _ 3 a^ h_3 h_2 _ n _ 3 + b^ h_3 j_2 _
n b^ h_3 h_2 _ n , + [ a_12_def ] a_12 & = & a^ h_3 j_2 _ n _ 3 b^ h_3 h_2 _
n -a^ h_3 h_2 _ n _ 3 b^ h_3j_2 _
n + & = & - _ 3 ^ 2 _ 2 ( h_n 3 ) ^2 ( k_2 ^ 2 - k_3 ^ 2 ) , + [ b_11_def ] b_11 & = & a^ h_2 j_1 _ n _ 2 a^ j_2 j_1 _ n _ 2 + b^ h_2 j_1 _
n b^ j_2 j_1 _ n , + [ b_12_def ] b_12 & = & a^ h_2 j_1 _ n _ 2 b^ j_2 j_1 _ n - a^ j_2 j_1 _ n _ 2 b^ h_2j_1 _
n + & = & - _ 1 ^ 2 _ 2 ( j_n1)^2 ( k_1 ^ 2 - k_2 ^ 2 ) . the function @xmath190 , @xmath191 , and @xmath192 , @xmath193 are obtained from @xmath194 , @xmath195 , and @xmath196 , @xmath197 , respectively , by replacing @xmath198 . in the last step in eqs .
( [ a_12_def ] ) and ( [ b_12_def ] ) we have used the wronskian determinant between bessel and hankel functions . finally , we have t_n 21 & = & a^ j_2 j_1 _ n _ 2 a^ j_2 j_1 _ n _ 2 + ( b^j_2j_1_n)^2 , + t_n 32 & = & a^ h_3 h_2 _ n _ 3 a^ h_3 h_2 _ n _ 3 + ( b^h_3h_2_n)^2 , d_n 21 & = & a^ h_2 j_1 _ n _ 2 a^ h_2 j_1 _ n _ 2 + ( b^ h_2 j_1 _
n ) ^2 , + d_n 32 & = & a^ h_3 j_2 _ n _ 3 a^ h_3 j_2 _ n _ 3 + ( b^ h_3 j_2 _
n ) ^2 , and [ a_hj_21_mu ] a^h_2j_1_n _ 2 & = & i _ 2 _ 1 ( _ 2 _ 1 h_n 2 j_n 1 - _ 1 _ 2 h_n 2 j_n 1 ) , + [ a_jj_21_mu ] a^j_2j_1_n _ 2 & = & i _ 2 _ 1 ( _ 2 _ 1 j_n 2 j_n 1 -_1 _ 2 j_n 2 j_n 1 ) , + [ a_hj_21_eps ] a_n _ 2^h_2 j_1 & = & i _ 2 _ 1 ( _ 2 _ 1 h_n 2 j_n 1 -_1 _ 2 h_n 2 j_n 1 ) , + [ a_hj_21 ]
a_n _ 2^j_2 j_1 & = & i _ 2 _ 1 ( _ 2_1 j_n 2 j_n1 -_1 _ 2 j_n 2 j_n 1 ) , + [ b_hj_21 ] b_n^h_2 j_1 & = & h_n2 j_n1 ( k_1 ^ 2-k_2 ^ 2 ) . whenever the combination @xmath199 and @xmath200 is involved in the superscript , the radius @xmath88 is implicit in the cylindrical functions .
for example , in eqs .
( [ a_hj_21_mu])([b_hj_21 ] ) we have & & z_n1z_n(_1 a_1 ) , z_n2z_n(_2 a_1 ) , + & & z_n1 , z_n2 .
the functions @xmath201 , @xmath202 , @xmath203 , @xmath204 , @xmath205 , @xmath206 , and @xmath207 are defined analogously , where we understand that the radius @xmath87 is implicit in all those functions .
of course , for the special case @xmath208 for all layers @xmath107 , these reflection coefficients simplify .
the reflection coefficients @xmath209 and @xmath210 can be obtained from @xmath211 and @xmath212 , respectively , by replacing @xmath213 .
note that the scattering coefficients @xmath212 as well as @xmath210 are dimensionless . however , the coefficients @xmath211 and @xmath209 are not , but the particular combinations @xmath214 are .
hinds and i.a .
hughes , j. phys .
phys . * 32 * , r119 ( 1999 ) ; r. folman , p. krger , j. schmiedmayer , j. denschlag , and c. henkel , adv . at .
* 48 * , 263 ( 2002 ) .
t. calarco , e.a .
hinds , d. jaksch , j. schmiedmayer , j.i .
cirac , and p. zoller , phys . rev .
a. 61 , 022304 ( 2000 ) .
johnson , phys .
rev . * 32 * , 97 ( 1928 ) ; h. nyquist , phys . rev . * 32 * , 110 ( 1928 ) .
jones , c.j .
vale , d. sahagun , b.v . hall , and e.a .
hinds , phys .
91 * , 080401 ( 2003 ) .
harber , j.m .
mcguirk , j.m .
obrecht , and e.a .
cornell , j. low . temp .
133 , 229 - 238 ( 2003 ) . c. henkel , s. ptting and m. wilkens , appl .
b * 69 * , 379 ( 1999 ) ; c. henkel and m. wilkens , europhys . lett . * 47 * , 414 ( 1999 ) .
l. knll , s. scheel , and d .-
welsch , _ qed in dispersing and absorbing media _ , in _ coherence and statistics of photons and atoms _ , ed .
j. peina ( wiley , new york , 2001 ) ; s. scheel , l. knll , d .-
welsch , and s.m .
barnett , phys .
a * 60 * , 1590 ( 1999 ) ; s. scheel , l. knll and d .-
welsch , phys .
a * 60 * , 4094 ( 1999 ) ; ho trung dung , l. knll and d .-
welsch , phys .
a * 62 * , 053804 ( 2000 ) .
agarwal , phys .
a * 11 * , 230 ( 1975 ) ; j.m .
wylie and j.e .
sipe , phys .
rev . a * 30 * , 1185 ( 1984 ) .
landau and e.m .
lifshitz , _ electrodynamics of continuous media _ ( pergamon press , oxford , 1960 ) .
l. onsager , phys . rev .
* 37 * , 405 ( 1931 ) ; * 38 * , 2265 ( 1931 ) ; w. eckhardt , opt
. commun . * 41 * , 305 ( 1982 ) ; phys .
a * 29 * , 1991 ( 1984 ) .
chew , _ waves and fields in inhomogeneous media _ ( ieee press , new york , 1990 ) .
jackson , _ classical electrodynamics _ , 2nd edn .
( wiley , new york , 1975 ) . e.m .
purcell , phys . rev . * 69 * , 681 ( 1946 ) .
tai , _ dyadic green functions in electromagnetic theory _
( ieee press , new york , 1993 ) .
z. xiang and y. lu , ieee trans .
microwave theory tech .
* 44 * , 614 ( 1996 ) . as pointed out in ref .
@xcite , this paper contains some critical mistakes .
the relation between the scattering coefficients and the green tensor in this paper is inconsistent .
li , m .- s .
leong , t .- s .
yeo , and p .- s .
kooi , j. of electromagnetic waves and appl .
* 14 * , 961 ( 2001 ) . | we derive a general expression for the spin - flip rate of an atom trapped near an arbitrary dielectric body and we apply this theory to the case of a @xmath0-layer cylindrical metal wire .
the spin flip lifetimes we calculate are compared with those expected for an atom near a metallic slab and with those measured by jones _
et al . _
above a 2-layer wire [ m.p.a .
jones , c.j .
vale , d. sahagun , b.v .
hall , and e.a .
hinds , phys .
rev .
lett . * 91 * , 080401 ( 2003 ) ] .
we investigate how the lifetime depends on the skin depth of the material and on the scaling of the dimensions .
this leads us to some conclusions about the design of integrated circuits for manipulating ultra - cold atoms ( atom chips ) . |
emerging technologies for computing promise to outperform conventional integrated circuits in computation bandwidth or speed , power consumption , manufacturing cost , or form factor @xcite .
however , razor - sharp focus on any one nascent technology and its benefits sometimes neglects serious limitations or discounts ongoing improvements in established approaches . to foster a richer context for evaluating emerging technologies
, we review limiting factors and salient trends in computing that determine what is achievable _ in principle _ and _ in practice_. several fundamental limits remain substantially loose , possibly indicating viable opportunities for emerging technologies . to clarify this uncertainty , we study
_ limits on fundamental limits_. * universal and general - purpose computers . * viewing _ clocks and watches _ as early computers , it is easy to see the importance of _ long - running calculations that can be repeated with high accuracy by mass - produced devices_. the significance of _ programmable _ digital computers became clear at least 200 years ago , as illustrated by _ jacquard looms _ in textile manufacturing .
however , the existence of _ universal computers _ that can efficiently simulate ( almost ) all other computing devices analog or digital was only articulated in the 1930s by church and turing ( turing excluded quantum physics when considering universality ) @xcite . efficiency was studied from a theoretical perspective at first , but strong demand in military applications in the 1940s lead turing and von neumann to develop detailed hardware architectures for universal computers turing s design ( pilot ace ) was more efficient , but von neumann s was easier to program .
the _ stored - program architecture _ made universal computers practical in the sense that a single computer design could be effective in many diverse applications .
such practical universality thrives @xmath0 in economies of scale in computer hardware and @xmath1 among extensive software stacks .
not surprisingly , the most sophisticated and commercially successful computer designs and components , such as intel and ibm cpus , were based on the von neumann s paradigm . the numerous uses and large markets of general - purpose chips , as well as the exact reproducibility of their results ,
justify the enormous capital investment in the design , verification and manufacturing of leading - edge integrated circuits . today general - purpose cpus power cloud server - farms and displace specialized ( but still universal ) mainframe processors in many supercomputers . emerging
universal computers based on field - programmable gate - arrays ( fpgas ) and general - purpose graphics processing units ( gpgpus ) outperform cpus in some cases , but their efficiencies remain complementary to those of cpus . the success of deterministic general - purpose computing manifests in the convergence of diverse functionalities in portable inexpensive smartphones . after steady improvement , general - purpose computing displaced entire industries ( newspapers , photography , etc ) and launched new applications ( video conferencing , gps navigation , online shopping , networked entertainment , etc ) @xcite .
application - specific integrated circuits ( asics ) streamline input - output and networking , or optimize functionalities previously performed by general - purpose hardware .
they speed up biomolecular simulation 100-fold @xcite and improve the efficiency of video decoding 500-fold @xcite , but require design effort with keen understanding of specific computations , impose high costs and financial risks , need markets where general - purpose computers lag behind , and often can not adapt to new algorithms .
recent techniques for _ customizable domain - specific computing _
@xcite offer better tradeoffs , while many applications favor the combination of _ general - purpose hardware and domain - specific software _ , including specialized programming languages @xcite such as erlang used in whatsapp .
* limits as aids to evaluating emerging technologies . * without sufficient history , we can not extrapolate _ scaling laws _ for emerging technologies , yet expectations run high . for example , new proposals for analog processors appear frequently ( as illustrated by adiabatic quantum computers ) , but fail to address concerns about analog computing , such as its limitations on scale , reliability , and long - running error - free computation .
general - purpose computers meet these requirements with digital integrated circuits ( ic ) and now command the electronics market . in comparison , quantum computers both digital and analog hold promise _ only in niche applications _ and do not offer _ faster general - purpose computing _ as they are no faster for _ sorting _ and other specific tasks @xcite .
in exaggerating the engineering impact of quantum computers , popular press missed this nuance .
but in scientific research , building quantum computers may help simulating quantum - chemical phenomena and reveal new fundamental limits .
sections [ sec : space ] and [ sec : conc ] discuss limits on emerging technologies .
* technology extrapolation versus fundamental limits . *
the scaling of commercial computing hardware regularly runs into formidable obstacles @xcite , but near - term technological advances often circumvent them .
the international technology roadmap for semiconductors ( itrs ) @xcite keeps track of such obstacles and possible solutions with a focus on frequently - revised consensus estimates .
for example , consensus estimates initially predicted 10 ghz cpus for the 45 nm technology node @xcite , versus the 3 - 4 ghz range seen in practice . in 2004 , the unrelated quantum information science and technology roadmap @xcite forecast 50 physical qubits by 2012 .
such optimism arose by assuming technological solutions long before they were developed and validated , and by overlooking important limits .
the authors of @xcite classify limits to device and interconnect as _ fundamental , material , device , circuit _ , and _
system limits_. these categories define the rows of table [ tab : limits ] , and the columns reflect sections of this paper where specific limits are examined for tightness .
engineering obstacles limit specific technologies and choices .
for example , a key bottleneck today is ic manufacture , which packs billions of transistors and wires in several @xmath2 of silicon with astronomically low defect rates .
layers of material are deposited on silicon and patterned with lasers , fabricating all circuit components simultaneously .
precision optics and photochemical processes ensure accuracy . * limits on manufacturing . *
no account of limits to computing is complete without the abbe diffraction limit : light with wavelength @xmath3 , traversing a medium with refractive index @xmath4 , and converging to a spot with angle @xmath5 ( perhaps , focused by a lens ) creates a spot with diameter @xmath6 , where @xmath7 is the numerical aperture .
@xmath8 reaches 1.4 for modern optics , so it would seem that semiconductor manufacturing is limited to feature sizes @xmath9 , hence arf lasers with 193 nm wavelength should not support photolithographic manufacturing of transistors with 65 nm features .
yet , they supported _ sub - wavelength lithography _ for the 45nm-22 nm technology nodes using _ asymmetric illumination _ and _ computational lithography _ @xcite .
here one starts with optical masks that look like the intended image , but when the image gets blurry , alter masks by gently shifting edges to improve the image , possibly giving up the semblance between the two .
clearly , some limits are formulated to be broken !
ten years ago , researchers demonstrated patterning of nanomaterials by live viruses @xcite .
known virions exceed 20 nm in diameter , whereas subwavelength lithography with 193nm - wavelength arf laser recently extended to 14 nm semiconductor manufacturing @xcite .
hence , viruses and microorganisms are no longer at the forefront of semiconductor manufacturing . extreme ultra - violet ( x - ray ) lasers have been energy - limited , but are improving .
their use requires changing refractive optics to reflective .
additional progress in _ multiple patterning _ and _ directed self - assembly _ promises to support photolithography beyond the 10 nm technology node .
* limits on individual interconnects . * despite the doubling of transistor density with moore s law @xcite , semiconductor integrated circuits ( ics ) would not work without fast and dense interconnects .
metallic wires can be either fast or dense , but not both at the same time smaller cross - section increases electrical resistance , while greater height or width increase parasitic capacitance with neighboring wires ( wire delay grows with @xmath10 ) . in 1995 ,
an intel researcher pointed out that _ on - chip interconnect scaling _ is the real limiter of high - performance ics @xcite .
the scaling of interconnect is also moderated by _
electron scattering against rough edges of metallic wires _
@xcite , inevitable with atomic - scale wires . hence
, ic interconnect stacks have evolved @xcite from four equal - pitch layers in 2000 to 16 layers with pitches varying by 32 times , including a large amount of dense ( thin ) wiring and fast ( thick ) wires used for global on - chip communication ( figure [ fig : wires ] ) .
aluminum and copper remain unrivaled for conventional interconnects and can be combined in short wires @xcite ; _ carbon - nanotube _ and _ spintronic _ interconnects are also evaluated in @xcite .
_ photonic waveguides _ and _ rf links
_ offer alternative ic interconnect @xcite , but obey fundamental limits derived from maxwell s equations , such as the _ maximum propagation speed of em waves _ @xcite .
i / o links are limited by the perimeter or surface area of a chip , whereas chip capacity grows with area or volume , respectively .
* limits on conventional transistors . * transistors are limited by their tiniest feature _ the width of the gate dielectric _ , which recently reached the size of several atoms ( figure [ fig : xtor - atoms ] ) , creating problems : @xmath0 a few missing atoms could alter transistor performance , @xmath1 manufacturing variation makes all transistors slightly different ( figure [ fig : xtor - field ] ) , @xmath11 electric current tends to leak through thin narrow dielectrics @xcite . instead of a _ thinner _
dielectric , transistors can be redesigned with _ wider _ dielectric layers @xcite that surround a _ fin shape _ ( figure [ fig : xtor - fin ] ) .
such configurations improve the control of electric field , reduce current densities and leakage , and diminish process variations .
each transistor can use several fins , extending transistor scaling by several generations .
semiconductor manufacturers adopted finfets for upcoming technology nodes .
one step further , in _ tunneling transistors _ @xcite a gate wraps around the channel to control tunnelling rate .
as mosfet transistors shrink , gate dielectric ( yellow ) thickness approaches several atoms ( 0.5 nm at the 22 nm technology node ) .
atomic spacing limits device density to 1 device / nm even for radical devices .
, width=294 ] as mosfet transistors shrink , the shape of electric field departs from basic rectilinear models , and level curves become disconnected .
atomic - level manufacturing variations , especially for dopant atoms , start affecting device parameters , making each transistor slightly different @xcite .
image credit : gold standard simulations .
, width=294 ] the evolution of metallic wire stacks from 1997 to 2010 by semiconductor technology nodes .
image credit : ibm research ( modified ) .
, width=302 ] finfet transistors possess much wider gate dielectric ( surrounding the fin shape ) than mosfet transistors and can use multiple fins .
, height=139 ] * limits on design effort . * in the 1980s , mead and conway formalized ic design using a regular grid , enabling automated layout through algorithms . but
resulting optimization problems remain hard , and heuristics are only good enough for practical use . besides frequent algorithmic improvements , each technology generation alters circuit physics and requires new cad software . the cost of design has doubled in a few years , becoming prohibitive for ics with limited market penetration @xcite .
emerging technologies , such as finfets and high - k dielectrics , circumvent known obstacles using forms of design optimization .
therefore , reasonably tight limits should account for potential future optimizations .
low - level technology enhancements , no matter how powerful , are often viewed as one - off improvements , in contrast to architectural redesigns that affect many processor generations . between technology enhancements and architectural redesigns
are global and local optimizations that alter `` the texture '' of ic design , such as _ logic restructuring , gate sizing and device parameter selection_. moore s law promises higher transistor densities , but some transistors are designed to be 32 times larger than others .
large gates consume greater power to drive long interconnects at acceptable speed and satisfy performance constraints . minimizing circuit area and power , subject to timing constraints ( by configuring each logic gate to a certain size , threshold voltage , etc ) ,
is a hard but increasingly important optimization with a large parameter space .
a recent convex optimization method @xcite saved 30% power in intel chips , and the impact of such improvements grows with circuit size .
many aspects of ic design are being improved , continually raising the bar for technologies that compete with cmos .
completing new ic designs , optimizing and verifying them requires great effort and continuing innovation , e.g. , the lack of scalable design automation is a limiting factor for analog ics @xcite . in 1999 , bottom - up analysis of digital ic technologies @xcite outlined design scaling up to self - contained modules with 50k standard cells ( each cell contains 1 - 3 logic gates ) , but further scaling was limited by global interconnect . in 2010 ,
physical separation of modules became less critical , as large - scale placement optimizations assumed greater responsibility for ic layout and learned to blend nearby modules @xcite . in a general trend ,
powerful design automation @xcite frees circuit engineers to focus on microarchitecture @xcite , but increasingly relies on algorithmic optimization . until recently
, this strategy suffered significant losses in performance @xcite and power @xcite compared to ideal designs , but has now became both successful and indispensable due to rapidly increasing complexity of digital and mixed - signal electronic systems .
hardware and software must now be co - designed and co - verified , with software efforts increasing at a faster rate .
_ platform - based design _ combines high - level design abstractions with effective reuse of components and functionalities in engineered systems @xcite .
customizable domain - specific computing @xcite and domain - specific programming languages @xcite offload specialization to software running on reusable hardware platforms .
in predicting the main obstacles to improving modern electronics , the international technology roadmap for semiconductors ( itrs ) highlights _ the management of system power and energy _ as the dominant grand challenge @xcite .
the faster the computation , the more energy it consumes , but actual power - performance tradeoffs depend on the physical scale .
while the itrs , by its charter , focuses on near - term projections and ic design techniques , fundamental limits reflect available energy resources , properties of the physical space , power - dissipation constraints , and energy waste .
a 1961 result by landauer @xcite shows that erasing one bit of information entails an energy loss @xmath12 ( _ thermodynamic threshold _ ) , where @xmath13 is the boltzmann constant and @xmath14 is the temperature in kelvin .
this principle was validated empirically in 2012 @xcite and seems to motivate _
reversible computing _
@xcite , where all input information is preserved , incurring additional costs .
formally speaking , zero - energy computation is prohibited by _ the energy - time form of the heisenberg uncertainty principle _ ( @xmath15 ) : faster computation requires greater energy @xcite .
however , recent work in applied superconductivity @xcite demonstrates `` highly exotic '' _ physically - reversible _ circuits operating at 4@xmath16k with energy dissipation below the thermodynamic threshold .
they apparently fail to scale to large sizes , run into other limits , and remain no more practical than `` mainstream '' superconducting circuits and refrigerated low - power cmos circuits .
technologies that implement _ quantum circuits _
@xcite can _ approximate _ reversible boolean computing , but currently do not scale to large sizes , are energy - inefficient at the system level , rely on fragile components , and require heavy fault - tolerance overhead @xcite .
conventional ics also do not help obtaining energy savings from reversible computing because they dissipate 30 - 60% of all energy in ( reversible ) wires and repeaters @xcite . at room temperature , landauer s
limit amounts to 2.85@xmath17 j a very small fraction of the total , given that modern ics dissipate 0.1 - 100 watts and contain @xmath18 logic gates . with the increasing dominance of interconnect ( section [ sec : space ] ) , more energy is spent on communication than on computation .
logically - reversible computing is important for reasons other than energy in cryptography , quantum information processing , etc @xcite . *
the end of cpu frequency scaling . * in 2004 , intel corp . abruptly cancelled a 4ghz cpu project because high power density required awkward cooling technologies .
other cpu manufacturers kept clock - frequencies in the 1 - 6ghz range , but also resorted to multicore cpus @xcite .
since dynamic circuit power grows with clock frequency and supply voltage squared @xcite , energy can be saved by distributing work among slower , lower - voltage parallel cpu cores _ if parallelization overhead is small_. * dark , darker , dim , gray silicon . * a companion trend to moore s law the dennard scaling theory @xcite shows how to keep power consumption of semiconductor ics constant while increasing their density . but
dennard scaling broke down ten years ago @xcite .
extrapolation of semiconductor scaling trends for cmos the dominant semiconductor technology for 20 years shows that the power consumption of transistors available in modern ics reduces more slowly than their size ( which is subject to moore s law ) @xcite . to ensure the performance envelope of transistors ,
chip power density must be limited , and a fraction of transistors must be kept dark at any given time .
modern cpus have not been able to use all their circuits at once , but this asymptotic effect termed the _ utilization wall _ @xcite will soon black out 99% of the chip , hence the term _ dark silicon _ and a reasoned reference to the apocalypse @xcite . saving power by slowing cpu cores down is termed _
dim silicon_. detailed studies of dark silicon @xcite show similar results . to this end
, executives from microsoft and ibm have recently proclaimed an end to the era of multicore microprocessors @xcite .
two related trends appeared earlier : @xmath0 increasingly large ic regions remain transistor - free to aid routing and physical synthesis , to accommodate power - ground networks , etc @xcite we call them _ darker silicon _ , @xmath1 increasingly many gates do not perform useful computation but reinforce long , weak interconnects @xcite or slow down wires that are too short call them _
gray silicon_. today , 50 - 80% of all gates in high - performance ics are repeaters .
* limits for power supply and cooling .
* data centers in the us consumed 2.2% of total u.s .
electricity in 2011 .
as powerplants take time to build , we can not sustain past trends of doubled power consumption per year .
it is possible to improve the efficiency of transmission lines ( using high - temperature superconductors @xcite ) and power conversion in datacenters , but the efficiency of on - chip power - networks may soon reach 80 - 90% .
modern ic power management includes _ clock _ and _ power gating _
@xcite , per - core voltage scaling @xcite , _ charge recovery _
@xcite and , in recent processors , a cpu core dedicated to power scheduling .
ic power consumption depends quadratically on supply voltage , which has decreased steadily for many years , but recently stabilized at 0.5 - 2v @xcite .
supply voltage typically exceeds the _ threshold voltage _ of field - effect transistors by a safety margin that ensures circuit reliability , fast operation and low leakage .
threshold voltage depends on the thickness of gate dielectric , which reached a practical limit of several atoms ( section [ sec : eng ] ) .
supply voltage is limited by around 200mv @xcite five times below current practice and simple circuits reach this limit . with slower operation , _ near- _ and _ sub - threshold circuits _ may consume 100 times less energy @xcite .
cooling technologies can improve too , but fundamental quantum limits bound the efficiency of heat removal @xcite .
the study in @xcite explores a general _ binary - logic switch model _ with binary states represented by two _ quantum wells separated by a potential barrier_. representing information by electric charge requires energy for binary switching and thus limits the logic - switching density , if a significant fraction of the chip can switch simultaneously . to circumvent this limit
, one can encode information in _ spin - states , photon polarizations , super - conducting currents _ , or _ magnetic flux _ , noting that these carriers have already been in commercial use .
spin - states are particularly attractive because they promise high - density nonvolatile storage @xcite and scalable interconnects @xcite .
more powerful limits are based on the amount of material in earth s crust ( where silicon is the second most common element after oxygen ) , on atomic spacing ( section [ sec : eng ] ) , radii and energies , bandgaps , as well as the wavelength of the electron .
we are currently using only a tiny fraction of earth s mass for computing , and yet various limits could be circumvented if new particles are discovered . beyond atomic physics ,
some limits rely on basic constants : the speed of light , the gravitational constant , the quantum ( planck ) scale , the boltzmann constant , etc .
lloyd @xcite , as well as kraus @xcite extend well - known bounds by bremermann and bekenstein , and give moore s law 150 and 600 years , respectively .
these results are too loose to obstruct the performance of practical computers .
in contrast , current consensus estimates from the itrs @xcite give moore s law only 10 - 20 years , due to both _
technological _ and _ economic _ considerations @xcite .
engineering limits for deployed technologies can often be circumvented , while first - principles limits on energy and power are very loose .
reasonably tight limits are rare .
* limits to parallelism .
* suppose we wish to compare a parallel and sequential computer built from the same units , to argue that a new parallel algorithm is many times faster than the best sequential algorithm ( the same reasoning applies to logic gates on an ic ) .
given @xmath19 parallel units and an algorithm that runs @xmath20 times faster on sufficiently large inputs , one can _ simulate _ the parallel system on the sequential system by dividing its time between @xmath19 computational slices .
since this simulation is roughly @xmath19 times slower , it runs @xmath21 times faster than the original sequential algorithm . if this algorithm was best possible , we have @xmath22 .
the bound is reasonably tight in practice for small @xmath19 and can be violated slightly since @xmath19 cpus include more cpu cache , but such violations do not justify parallel algorithms one could instead buy / build one cpu with a larger cache .
such linear speedup is optimistically assumed for _ the parallelizable component _ in the 1988 gustafson s law that suggests scaling the number of processors with input size ( as illustrated by instantaneous web search queries over massive data sets ) @xcite . also in 1988 ,
fisher @xcite employed _ asymptotic runtime estimates _ instead of numerical limits and avoided the breakdown into parallel and sequential runtime components , assumed in amdahl s @xcite and gustafson s laws @xcite .
asymptotic estimates neglect leading constants and offer a powerful way to capture nonlinear phenomena occurring at large scale .
fisher @xcite assumes a sequential computation with @xmath23 elementary steps for input of size @xmath24 , and limits the performance of its parallel variants that can use an unbounded @xmath25-dimensional grid of finite - size computing units ( electrical switches on a semiconductor chip , logic gates , cpu cores , etc ) communicating at a finite speed , say , bounded by the speed of light .
we highlight only one aspect of this four - page work the parallel computation requires @xmath26{t(n)})$ ] steps .
this result undermines the @xmath19-fold speedup assumed in gustafson s law for @xmath19 processors on appropriately sized input data @xcite . a more realistic speedup from @xmath27 to @xmath28
can be achieved in an abstract model of computation for matrix multiplication and fast fourier transforms .
but not in physical space @xcite . surprising as it may seem , after reviewing many loose limits to computation , we have identified a reasonably tight limit ( the impact of i / o a major bottleneck today is also covered in@xcite ) .
indeed , many parallel computations today ( excluding multimedia processing and web search ) are limited by several forms of communication and synchronization , including network and storage access .
the billions of logic gates and memory elements in modern ics are linked by up to 16 levels of wires ( figure [ fig : wires ] ) , longer wires are segmented by repeaters .
most of the physical volume and circuit delay are attributed to interconnect @xcite .
this is relatively new , as gate delays were dominant until 2000 @xcite , but wires get slower relative to gates at each new technology node .
this uneven scaling has compounded in ways that would surprise turing and von neumann a single clock cycle is now far too short for a signal to cross the entire chip , and even the distance covered in 200 ps ( 5 ghz ) at light - speed is close to chip size . yet , most electrical engineers and computer scientists continue to focus on gates .
* implications to 3d ics and other emerging technologies . * the promise of 3d integration for improving ic performance can be contrasted with technical obstructions to its industry adoption . to derive limits on possible improvement
, we use the result from @xcite sensitive to the dimension of the physical space : a sequential computation with @xmath23 steps requires @xmath29{t(n)})$ ] steps in 2d and @xmath30{t(n)})$ ] in 3d . letting @xmath31{t(n)}$ ] , shows that 3d integration asymptotically reduces @xmath32 to @xmath33 a significant but not dramatic speedup .
this speedup requires an unbounded number of 2d device layers , otherwise there is no asymptotic speedup @xcite . for 3d ics with 2 - 3 layers ,
the main benefits of 3d ic integration today are in improving manufacturing yield , improving i / o bandwidth , and combining 2d ics that are optimized for random logic , dense memory , fpga , analog , mems , etc .
ultra - high density cmos logic ics with _ monolithic _
3d integration @xcite suffer higher routing congestion than traditional 2d ics .
emerging technologies promise to improve device parameters , but often remain limited by scale , faults , and interconnect , e.g. , _ quantum dots _
enable terahertz switching but hamper nonlocal communication @xcite .
cnt - fets @xcite leverage extraordinary carrier mobility in semiconducting carbon nanotubes to use interconnect more efficiently by improving drive strength , while reducing supply voltage .
emerging interconnects include _ silicon photonics _ , shown by intel in 2013 @xcite as a 100gb / s replacement of copper cables connecting adjacent chips .
it promises to reduce power consumption and form factor .
quantum physics alters the nature of communication with einstein s `` spooky action at a distance '' facilitated by entanglement @xcite .
however , the flows of information and entropy are subject to quantum limits @xcite .
several quantum algorithms run asymptotically faster than best conventional algorithms @xcite , but fault - tolerance overhead offsets their potential benefits in practice , and empirical evidence of quantum speedups has not been compelling so far @xcite .
several stages in the development of quantum information processing remain challenging @xcite , and the surprising difficulty of _ scaling up _ reliable quantum computation could stem from limits on _ communication _ and _ entropy _ @xcite .
in contrast , lloyd @xcite notes that _ individual _ quantum devices now approach energy limits for switching , whereas nonquantum devices remain orders of magnitude away .
this suggests an obstacle to simulating quantum physics on conventional computers ( abstract models aside ) . in terms of computational complexity
though , quantum computers _ can not _ attain significant advantage for many problem types @xcite . such lack of _ consistent general - purpose speedup _ limits the benefits of several emerging technologies in mature applications with diverse algorithmic steps , e.g. , computer - aided design and web search .
accelerating one step usually does not greatly speed up the entire application , as noted by amdahl in 1967 @xcite .
figuratively speaking , _
the most successful computers are designed for the decathlon , rather than for sprint only_.
section [ sec : space ] enabled tighter limits by neglecting energy and using asymptotic rather than numeric bounds a more abstract model focuses on the impact of scale , and recurring trends quickly overtake one - off device - specific effects .
next , we neglect spatial effects and focus on the nature of computation in an abstract model ( used by software engineers ) that represents computation by elementary steps with input - independent runtimes .
such limits survive many improvements in computer technologies , and are often stronger for specific problems .
for example , the best - known algorithms for multiplying large numbers are only slightly slower than reading the input ( an obvious speed limit ) , but only in the asymptotic sense for numbers with @xmath341000 bits , those algorithms lag behind simpler algorithms in actual performance . to focus on
what matters , we now do not just track asymptotic worst - case complexity of best algorithms for a given problem , but merely distinguish _ polynomial _ asymptotic growth from _
exponential_. limits formulated in such crude terms ( unsolvability in polynomial time _ on any computer _ ) are powerful @xcite : the hardness of number - factoring underpins internet commerce , while the p@xmath35np conjecture explains the lack of satisfactory , scalable solutions to important algorithmic problems , e.g. , in optimization and verification of ic designs @xcite .
a similar conjecture p@xmath35nc seeks to explain why many algorithmic problems that can be _ solved _ efficiently have not _ parallelized _ efficiently @xcite .
most of these limits have not been proven .
some can be circumvented by using radically different physics , e.g. , quantum computers solve number factoring in polynomial time ( in theory ) .
but quantum computation does not affect p@xmath35np @xcite .
the lack of proofs , despite heavy empirical evidence , requires faith and is an important limitation of many nonphysical limits to computing .
this faith is not universally shared donald knuth argues that p = np would not contradict anything we know today .
a rare _ proven _ result by turing ( also invulnerable to quantum physics ) states that checking if a given program ever halts is _ undecidable _ : no algorithm solves this problem in all cases regardless of runtime . yet
, software developers solve this problem during peer code reviews , and computer science teachers when grading exams in programming courses .
_ worst - case analysis _ is another limitation of nonphysical limits to computing , but suggests potential gains through approximation and specialization .
for some np - hard optimization problems , such as the _ euclidean travelling salesman problem _ ( euctsp ) , polynomial - time approximations exist , but in other cases , such as the _ maximum clique problem _ , accurate approximation is as hard as finding optimal solutions @xcite .
for some important problems and algorithms , such as the simplex algorithm for _ linear programming _ , few inputs lead to exponential runtime , and minute perturbations reduce runtime to polynomial @xcite .
the death march of moore s law @xcite invites discussions of fundamental limits and alternatives to silicon semiconductors @xcite .
near - term constraints invariably tie to _ costs _ and _ capital _ , but are explained away by new markets for electronics , increasing earth population , and growing world economy @xcite .
such economic pressures emphasize the value of _ computational universality _ and broad applicability of ic architectures to solve multiple tasks under conventional environmental conditions . in a likely scenario ,
only cpus , gpus , fpgas and dense memory ics will remain viable at the end of moore s law , while specialized circuits will be manufactured with less advanced technologies . indeed , memory chips have lead moore scaling by leveraging their simpler structure , modest interconnect , and more controllable manufacturing , but their scaling is slowing down @xcite . the decelerated scaling of cmos ics still outperforms the scaling of the most viable emerging technologies
. empirical scaling laws describing the evolution of computing are well - known @xcite .
in addition to moore s law , dennard scaling , as well as amdahl s and gustafson s laws reviewed earlier , metcalfe s law @xcite states that the value of a computer network , such as the internet or facebook , scales as the number of user - to - user connections that can be formed .
grosch s law @xcite ties @xmath19-fold improvements in computer performance to @xmath36-fold cost increases ( in equivalent units ) .
applying it in reverse , we can estimate acceptable performance of cheaper computers .
but such laws only capture _ ongoing scaling _ and will break down in the future . the _ roadmapping process _ represented by the international technology roadmap for semiconductors ( itrs )
@xcite relies on consensus estimates and works around engineering obstacles .
it tracks improvements in materials and tools , collects best practices and outlines promising design strategies . as suggested in @xcite , it can be enriched by analysis of limits .
we additionally focus on how closely such limits can be approached . aside from historical `` wrong turns '' recalled in sections [ sec : eng ] and [ sec : energy ] , we find interesting effects when examining the tightness of individual limits . while energy - time limits are most critical in computer design @xcite , space - time limits appear tighter @xcite and capture bottlenecks formed by interconnect and communication .
they suggest optimizing gate locations and sizes , and placing gates in three dimensions .
one can also adapt algorithms to spatial embeddings @xcite and seek space - time limits .
but the gap between current technologies and energy - time limits hints at greater rewards . _ charge recovery _
@xcite , _ power management _
@xcite , voltage scaling @xcite , and _ near - threshold computing _
@xcite reduce energy waste .
optimizing algorithms and circuits simultaneously for energy and spatial embedding @xcite gives biological systems an edge ( from the 1d worm _ c. elegans _ with 302 neurons to the 3d human brain with 86 billion neurons ) @xcite .
yet , using mass - energy to compute can be a veritable _ nuclear option_. in a 1959 talk , which predated moore s law , richard feynman suggested that there was `` plenty of room at the bottom , '' forecasting the miniaturization of electronics . today
, with relatively little physical room left , _ there is plenty of _ energy _ at the bottom_. if this energy is tapped for computing , how can resulting heat be removed ? recycling heat into mass or electricity seems ruled out by limits to energy conversion and the acceptable thermal envelope .
technology - specific limits for modern computers tend to express tradeoffs , especially for systems with conflicting performance parameters and properties @xcite .
little is known about limits on _
design technologies_. given that large - scale complex systems are often designed and implemented hierarchically @xcite with multiple levels of abstraction , it would be valuable to capture losses incurred at abstraction boundaries and between levels of design hierarchies .
it is common to estimate resources required for a subsystem and then implement the subsystem to satisfy resource budgets .
underestimation is avoided because it leads to failures , but overestimation results in overdesign .
inaccuracies in estimation and physical modeling also lead to losses during optimization , especially in the presence of uncertainty .
clarifying engineering limits gives hope to circumvent them .
technology - agnostic limits look simple and have had significant impact in practice , for example aaronson explains why np - hardness is unlikely to be circumvented by through physics @xcite .
limits to _ parallel computation _ became prominent after cpu speed levelled off ten years ago .
they suggest using faster interconnect @xcite , local computation that reduces communication @xcite , time - division multiplexing of logic @xcite , architectural and algorithmic techniques @xcite , solving larger problem instances , and altering applications to embrace parallelism @xcite .
john gustafson advocates a _ natural selection _ : the survival of applications fittest for parallelism . in another twist ,
the performance and power consumption of industry - scale distributed systems is often described by probability distributions , rather than single numbers @xcite , making it harder to even formulate appropriate limits .
we also can not yet formulate fundamental limits related to the complexity of the software - development effort , the efficiency of cpu caches @xcite , and computational requirements of incremental functional verification , but we have noticed that many known limits are either loose or can be circumvented , leading to _
secondary limits_. to wit , the @xmath37 limit is worded in terms of worst - case rather than average - case performance , and has not been proven despite heavy evidence .
researchers have ruled out entire categories of proof techniques as insufficient to complete such a proof @xcite . while esoteric , such _ tertiary limits _ can be effective in practice
in august 2010 , they helped researchers quickly invalidate vinay deolalikar s highly - technical attempt at proving @xmath37 . on the other hand , the correctness of lengthy proofs for some key results could not be established with acceptable level of certainty by reviewers , prompting efforts in verifying mathematics by computation @xcite . in summary
, we have reviewed what is known about limits to computation , including existential challenges arising in the sciences , design and optimization challenges arising in engineering , as well as current state of the art .
these categories are closely linked due to the rapid pace of technology development . when a specific limit is approached and obstructs progress , understanding its assumptions is a key to circumventing it .
some limits are hopelessly loose and can be ignored , while other limits remain conjectured based on empirical evidence and may be very difficult to establish rigorously .
such _ limits on limits to computation _ deserve further study .
* acknowledgments .
* this work was supported in part by the semiconductor research corporation ( src ) task 2264.001 ( funded by intel and ibm ) , us airforce research laboratory award fa8750 - 11 - 2 - 0043 , and us national science foundation ( nsf ) award 1162087 .
hameed , r. , qadeer , w. , wachs , m. , azizi , o. , solomatnikov , a. , lee , b. c. , richardson , s. , kozyrakis , c. , horowitz , m. understanding sources of ineffciency in general - purpose chips , _ commun .
54(10 ) : 85 - 93 october ( 2011 ) .
davis , j. a. , venkatesan , r. , kaloyeros , a. , beylansky , m. , souri , s. j. , banerjee , k. , saraswat , k. c. , rahman , a. , reif , r. , meindl , j. d. interconnect limits on gigascale integration ( gsi ) in the 21st century , _ proc . ieee _
89(3):305 - 324 ( 2001 ) .
hisamoto , d. , lee , w .- c . ,
kedzierski , j. , takeuchi , h. , asano , k. , kuo , c. , anderson , e. , king , t .- j .
, bokor , j. , hu , c. finfet - a self - aligned double - gate mosfet scalable to 20 nm , _ ieee trans . on electron devices _ 47(12):2320 - 2325 ( 2002 ) .
brut , a. , arakelyan , a. , petrosyan , a. , ciliberto , s. , dillenschneider r. , and lutz , e. experimental verification of landauer s principle linking information and thermodynamics , _ nature _ 483 : 187 - 189 ( 2012 )
. monroe , c. , raussendorf , r. , ruthven , a. , brown , k. r. , maunz , p. , duan , l .- m . , and kim , j. large - scale modular quantum - computer architecture with atomic memory and photonic interconnects , _ phys . rev .
a _ 89 , 022317 .
borkar , s. thousand - core chips : a technology perspective , in _ proc .
design automation conf .
_ ( dac ) ) 746 - 749 ( 2007 ) .
rabaey , j. m. , chandrakasan , a. , nikolic , b. digital integrated circuits a design perspective , _ pearson education , inc _ ( 2003 ) .
bohr , m. a 30 year retrospective on dennard s mosfet scaling paper , _ ieee solid - state circuits society newsletter _
12(1 ) : 11 - 13 ( 2007 ) .
taylor , m. b. is dark silicon useful ?
harnessing the four horsemen of the coming dark silicon apocalypse , in _ proc .
design automation conf . _
( dac ) 1131 - 1136 ( 2012 ) .
dreslinski , r. g. , wieckowski , m. , blaauw , d. , sylvester , d. , mudge , t. near - threshold computing : reclaiming moore s law through energy efficient integrated circuits , _ proc .
98(2 ) : 253 - 266 ( 2010 ) .
wolf , s. a. , awschalom , d. d. , buhrman , r. a. , daughton , j. m. , von molnr , s. , roukes , m. l. , chtchelkanova , a. y. , treger , d. m. , spintronics : a spin - based electronics vision for the future , _ science _ 294:1488 - 1494 ( 2001 ) .
lee , y .- j . ,
morrow , p. , lim , s. k. , `` ultra high density logic designs using transistor - level monolithic 3d integration , '' in _ proc .
. computer - aided design of integrated circuits _ _ iccad _ 2012 : 539 - 546 .
dror , r. o. , grossman , j. p. , mackenzie , k. m. , towles , b. , chow , e. , salmon , j. k. , young , c. , bank , j. a. , batson , b. , shaw , d. e. , kuskin , j. , larson , r. h. , moraes , m. a. , shaw , d. e. overcoming communication latency barriers in massively parallel scientific computation , _ ieee micro _
31(3 ) : 8 - 19 ( 2011 ) .
barroso , l. a. , clidaras , j. , hlzle , u. the datacenter as a computer : an introduction to the design of warehouse - scale machines , 2nd ed . : synthesis lectures on computer architecture ( _ morgan & claypool publishers _ 2013 ) . | an indispensable part of our lives , _ computing _ has also become essential to industries and governments .
steady improvements in computer hardware have been supported by periodic doubling of transistor densities in integrated circuits over the last fifty years .
such _
moore scaling _ now requires increasingly heroic efforts , stimulating research in alternative hardware and stirring controversy . to help evaluate emerging technologies and
enrich our understanding of integrated - circuit scaling , we review fundamental limits to computation : in manufacturing , energy , physical space , design and verification effort , and algorithms . to outline what is achievable _ in principle _ and _ in practice _ , we recall how some limits were circumvented , compare loose and tight limits .
we also point out that engineering difficulties encountered by emerging technologies may indicate yet - unknown limits . |
upcoming ground - based laser interferometric gravitational wave detectors such as advanced ligo @xcite , advanced virgo @xcite , geo - hf @xcite and lcgt @xcite are expected to be limited by thermal noise in their most sensitive frequency band around @xmath8hz .
currently , strategies for detectors beyond the 2nd generation , such as the einstein telescope @xcite , are being developed aiming for a ten times better sensitivity . following
levin s approach @xcite mechanical dissipation located at a mirror front surface contributes more than the same dissipation within the substrate .
the dominant contribution to thermal noise in the detection band of advanced detectors arises from multilayer coatings made of tantala and silica , used for highly reflective test mass mirrors at a laser wavelength of @xmath3 nm @xcite .
hence , a reduction of the amount of mechanically lossy coating material is expected to improve the coating brownian thermal noise of a test mass mirror .
one approach being considered is based on resonant waveguide grating structures as a substitute for commonly used multilayer coatings . in particular , broadband waveguide grating structures under normal incidence have been proposed as an alternative concept for test mass mirrors , providing high reflectivity while having significantly less @xcite , or perhaps even no @xcite , additional coating material .
both concepts have been successfully realized and tested in bench - top cavity experiments where the highest reflectivities seen so far are @xmath9 at @xmath3 nm @xcite and @xmath10 at @xmath11 nm @xcite . here ,
we report on the next step towards the implementation of waveguide grating mirrors in large - scale gravitational wave detectors , namely their usage as a coupling mirror in a fully suspended low - noise prototype environment .
therefore , the glasgow prototype facility @xcite was commissioned to incorporate a custom made waveguide grating mirror in a @xmath0 m linear fabry - perot cavity .
we could demonstrate stable operation of the waveguide cavity with a measured finesse of about @xmath1 .
the waveguide grating design was based on a tantala layer and silica substrate having an additional etch stop layer of al@xmath5o@xmath6 , which allowed us to define the grating depth accurately in the fabrication process .
its reflectivity was determined to be @xmath12 , which is the highest reflectivity of such a device reported at a laser wavelength of @xmath3 nm so far .
a nanostructured waveguide layer can provide resonant excitation of an incoupling light field , which results in anomalies of its reflection coefficient based on leaky waveguide modes as first found by hessel and oliner @xcite and later experimentally confirmed for a first and zeroth diffraction order @xcite .
this property is , for instance , widely investigated for use in narrowband filter applications @xcite
. however , broadband designs under normal incidence as investigated in @xcite are favourable for the purpose of highly reflective surface mirrors in high - precision interferometric experiments , since they are less sensitive to parameter deviations .
the basic principle of a waveguide grating mirror under normal incidence is shown in fig . [
fig : architecture](a ) using a ray picture @xcite . .
the outcoupling at the grating structure is a constructive interference if all grating parameter are designed properly , hence , providing @xmath13 reflectivity .
( b ) equivalent waveguide grating architecture realized in this work , including an al@xmath5o@xmath6 layer as etch stop in order to define the grating depth in the fabrication process.,width=453 ] it follows the idea of having a single nanostructured layer with high index of refraction @xmath14 on a substrate having a lower index of refraction of @xmath15 .
incident light of wavelength @xmath16 can then be resonantly enhanced for subwavelength grating structures if the grating parameters , namely the grating period @xmath17 , waveguide layer thickness @xmath18 , grating depth @xmath19 and fill factor @xmath20 ( ratio of ridge width @xmath21 and grating period @xmath17 ) , are designed properly . from the grating equation
one can derive a lower boundary for the grating period that permits the propagation of first diffraction orders in the waveguide layer and an upper boundary that allows guidance of light due to total internal reflection at the interface of high and low index of refraction materials @xcite . the same boundaries simultaneously prohibit higher diffraction orders in air ( @xmath22 ) and the waveguide layer ( @xmath23 ) . as a result
, the grating period @xmath17 is restricted to a range of @xmath24 for tantala ( ta@xmath5o@xmath25 ) with @xmath26 and fused silica ( sio@xmath5 ) with @xmath27 at a wavelength of @xmath28 nm .
this restriction is illustrated in fig .
[ fig : gratingperiod](a ) by means of rigorous coupled wave analysis ( rcwa ) calculations @xcite for a tantala grating structure and te - polarized light ( electric field vector parallel to the grating ridges ) .
the white lines mark the upper and lower boundary for the grating period with respect to the indices of refraction for the substrate and waveguide layer material , respectively . ,
@xmath29 nm , @xmath30 , te - polarization ) into a material with varying index of refraction @xmath15 , illustrating the range of the grating period that allows for resonant excitation as predicted by eq . (
[ eq : range ] ) .
( b ) if an etch stop layer with index of refraction @xmath31 is implemented , coupling into the waveguide occurs for @xmath32 independent of its thickness @xmath33.,width=453 ] for the waveguide grating described in this article , we have implemented an additional thin layer of al@xmath5o@xmath6 ( see fig .
[ fig : architecture](b ) ) , which was used as an etch stop ( there is a high contrast of etching rates between al@xmath5o@xmath6 and ta@xmath5o@xmath25 ) to define a particular grating depth in the fabrication process .
its index of refraction of @xmath31 is high enough to allow first order coupling to the waveguide layer if @xmath32 holds for the grating period independent of the etch stop layer thickness @xmath33 ( see fig . [
fig : gratingperiod](b ) ) . in particular for a thickness of only a few nanometers there is no significant influence on the coupling efficiency and consequently the waveguide grating s optical properties in comparison with the conventional design ( @xmath34 nm in fig .
[ fig : gratingperiod](b ) ) .
the starting point for the fabrication was a standard @xmath35inch fused silica mask blank coated by a layer system of tantala ( @xmath36 nm ) , al@xmath5o@xmath6 ( @xmath37 nm ) and tantala ( @xmath29 nm ) .
a top chromium ( cr ) layer of @xmath38 nm thickness was attached onto the layer system , serving as the mask during tantala etching .
this mask was realized by spin coating the whole sample with an electron beam sensitive resist and applying electron beam lithography for an area of @xmath39 mm aiming at a grating period of @xmath40 nm , which satisfies @xmath41 , and a fill factor of about @xmath42 . after
resist development the chromium layer was structured by utilizing an inductively coupled plasma ( icp ) dry - etching process .
finally the binary chromium mask was transferred only into the upper tantala layer by means of an anisotropic reactive ion beam etching ( ribe ) process supported by the high etching contrast between tantala and al@xmath5o@xmath6 . a scanning electron microscope ( sem )
cross - sectional image of a fabricated structure from this process is shown in fig .
[ fig : semdesign](a ) , which is based on the same layer system and etching process but different fill factor as the one used in the experiment . .
the white lines mark an area of reflectivity @xmath43.,width=453 ] the sample preparation for sem characterization was done by using a focused ion beam ( fib ) which requires covering the grating spot of interest by a platinum ( pt ) layer ( see fig .
[ fig : semdesign](a ) ) .
this image also shows the residual chromium mask , which was removed prior to sample application . corresponding to the values for the grating parameters given in fig .
[ fig : semdesign](a ) , rcwa was used to predict the reflectivity for te - polarized light and a fill factor of @xmath44 ( see fig .
[ fig : semdesign](b ) ) .
these results indicate that reflectivities higher @xmath45 are feasible , within parameter uncertainties of a few nanometers that are difficult to resolve on the basis of the sem image .
the most crucial parameter here is the tantala layer thickness @xmath18 . a variation of this layer thickness by @xmath46 nm leads to an absolute decrease in reflectivity of @xmath47 , since the area of high reflectivity in fig .
[ fig : semdesign](b ) shifts to larger grating periods with smaller waveguide layer thickness ( not shown here ) . we would like to note , that the etch stop layer design in principle offers the possibility to reduce the grating depth @xmath19 of an already fabricated grating sample without affecting the other parameters by means of an additional etching step , which has not been done yet for our investigated device . according to fig .
[ fig : semdesign](b ) this could efficiently be used in conjunction with a slightly smaller grating period ( or in case of a tantala layer thickness @xmath48 nm ) , which enables to cross the area of high reflectivity with decreasing grating depth .
figure [ fig : jif](a ) shows a schematic overview of the section of the glasgow 10 m prototype used for the work described in this article .
m. the cavity was stabilized using the pound - drever - hall scheme also depicted at the laser bench .
( b ) intermediate and lower stage of the triple suspension system used for the cavity mirrors . behind the main test mass a second triple suspension carries the so - called reaction mass , which is used to act on the main test mass .
( c ) test mass with the waveguide grating mirror ( area of @xmath49 mm ) attached.,width=453 ] a nd : yag laser at a wavelength of @xmath3 nm is spatially filtered by a single - mode fiber before being injected to the vacuum system and guided to the waveguide cavity via a beam splitter ( bs ) and a steering mirror ( sm ) that both are realized as double suspensions .
the cavity mirrors are suspended as triple pendulums ( see fig . [
fig : jif](b ) ) based on the geo600 suspension design @xcite . while the end mirror was a conventional multilayer mirror with a nominal power reflectivity of @xmath50 the coupling to the cavity was realized with the waveguide grating mirror having an area of @xmath39 mm ( see fig . [
fig : jif](c ) ) .
their separation was @xmath51 m , yielding a free spectral range of @xmath52mhz with @xmath53 being the speed of light .
the radius of curvature of the end mirror ( @xmath54 m ) defines the beam waist diameter on the plane waveguide grating to be @xmath55 mm , which is at least three times smaller than the grating .
therefore , we assume any power losses from beam clipping to be negligible . for all measurements the vacuum system had been evacuated to @xmath56 mbar in order to suppress acoustic noise and residual gas pressure noise .
the cavity could be stabilized on resonance using the pound - drever - hall scheme @xcite .
a phase modulation at @xmath57mhz was imprinted on to the incident light via an electro - optical - modulator ( eom ) , which is equivalent to sidebands at @xmath58mhz from the first fsr .
the reflected light was detected with the photodiode pd@xmath59 and electronically demodulated resulting in an bipolar error signal with zero crossing at the cavity resonance , which was fed back to the laser frequency . in fig .
[ fig : fsr ] the reflected light power ( red trace ) and the error - signal ( green trace ) are shown , while the laser frequency was ramped .
the reflected signal for the stabilized cavity ( blue trace ) shows a visibility of @xmath60 , which is only a lower value due to power present in higher order modes and signal - sidebands . and
pd@xmath59 , respectively .
the reflected signal for a stabilized cavity ( blue trace ) indicates a visibility of @xmath61.,width=453 ] the cavity finesse was already high enough to exhibit the dynamical effect of ringing for a sweep through resonance as shown in fig .
[ fig : ring ] for different mirror velocities @xmath62 .
( red trace ) .
theoretical results for the transmitted power @xmath63 are based on eq .
( [ eq : trans ] ) with @xmath64 ( black line ) , which corresponds to a cavity finesse of 790 . hence ,
a lower boundary for the waveguide grating power reflectivity is @xmath65.,width=453 ] therefore , the cavity end mirror was pushed longitudinally via its reaction mass ( magnet - coil - actuators ) .
the end mirror s velocity with respect to the coupling mirror was determined via @xmath66 where @xmath67 is the sideband frequency at @xmath68mhz and @xmath69 is the separation in time of the two sideband signals around resonance .
the theoretical model for the light field dynamics of a swept cavity used here ( see e.g. @xcite ) is based on a constantly moving end mirror with a velocity @xmath70 , giving a time dependent cavity length of @xmath71 . the transmitted field @xmath72 consists of a number of partial beams @xmath73 that have undergone @xmath74 cavity transits .
their sum is written as @xmath75 where @xmath76 is the incident light field and @xmath77 and @xmath78 denote the amplitude reflectivity and transmissivity of the coupling and end mirror , respectively .
the round - trip phase for the @xmath79-th partial field can be approximated to @xmath80 with the assumption that the cavity length change with each round - trip is negligible for calculating the round - trip time @xmath81 .
the first term in eq .
( [ eq : phi ] ) describes a cavity where at each time @xmath82 an equilibrium for the intra - cavity field is reached thus giving the well - known airy peaks .
the second term does account for additional phases due to the non - zero round - trip time of light with respect to the mirror motion .
the theoretical results in fig . [ fig : ring ] are based on a product of amplitude reflectivities of @xmath83 , which yields a cavity finesse of @xmath84 .
the error given here arises from the uncertainty of the cavity length of @xmath85 m and measured time separation @xmath69 of @xmath86 , which both determine the mirror velocity via eq .
( [ eq : v ] ) .
if we assume a perfect end mirror @xmath87 and zero round - trip loss a lower value for the waveguide grating power reflectivity is @xmath88 .
this value is in good agreement with the numerical prediction shown in fig .
[ fig : semdesign](b ) ( @xmath89 nm , @xmath29 nm ) thus supporting the principle of the waveguide grating architecture investigated here .
we have demonstrated stable operation of a fully suspended @xmath0 m cavity incorporating a waveguide grating as coupling mirror , which is a key step towards the application of waveguide mirrors as test masses of future gravitational wave detectors .
a cavity finesse of about @xmath1 was determined from the deformation of airy peaks ( i.e. ringing effect ) for a sweep over the cavity resonance .
the corresponding waveguide grating reflectivity of @xmath12 is , to the best of our knowledge , the highest reflectivity ever realized for a resonant waveguide grating at a laser wavelength of @xmath3 nm . the result is in good agreement with rigorous simulations , thus supporting the principle of the investigated design , which in principle enables accurate control of the grating depth via further etching . whether the optical and mechanical properties of an additional thin etch stop layer made of al@xmath5o@xmath6 can meet the strict requirements for future detector test masses need to be further investigated . in future
we also plan to characterize in more detail the noise performance of the suspended waveguide mirror .
a special focus will be put on investigating the level of any potential phase noise coupling from side motion @xcite of the waveguide mirror .
this work was supported by the deutsche forschungsgemeinschaft within the sonderforschungsbereich tr7 , the excellence cluster quest and by the university of glasgow and the science and technology facilities council ( stfc ) .
we also like to acknowledge the support of the imprs on gravitational wave astronomy and the european commission under the framework programme 7 ( grant agreement 211743 ) . | we report on the first demonstration of a fully suspended @xmath0 m fabry - perot cavity incorporating a waveguide grating as the coupling mirror .
the cavity was kept on resonance by reading out the length fluctuations via the pound - drever - hall method and employing feedback to the laser frequency . from the achieved finesse of @xmath1 the grating reflectivity was determined to exceed @xmath2 at the laser wavelength of @xmath3 nm , which is in good agreement with rigorous simulations .
our waveguide grating design was based on tantala and fused silica and included a @xmath4 nm thin etch stop layer made of al@xmath5o@xmath6 that allowed us to define the grating depth accurately during the fabrication process .
demonstrating stable operation of a waveguide grating featuring high reflectivity in a suspended low - noise cavity , our work paves the way for the potential application of waveguide gratings as mirrors in high - precision interferometry , for instance in future gravitational wave observatories . 99 g. m. harry ( for the ligo scientific collaboration ) , `` advanced ligo : the next generation of gravitational wave detectors , '' class .
quantum grav .
* 27 , * 084006 ( 2010 ) .
f. acernese et al . , `` advanced virgo baseline design , '' vir-027a-09 ( https://tds.ego-gw.it/ql/?c=6589 ) .
b. willke et al .
, `` the geo - hf project , '' class .
quantum grav .
* 23 , * s207s214 ( 2006 ) .
k. kuroda and the lcgt collaboration,``status of lcgt , '' class .
quantum grav .
* 27 , * 084004 ( 2010 ) .
m. punturo et al .
`` the einstein telescope : a third - generation gravitational wave observatory , '' class . and quantum grav .
, * 27 * , 194002 ( 2010 ) .
s. hild et al . `` a xylophone configuration for a third - generation gravitational wave detector , '' class . and quantum grav .
, * 27 * , 015003 ( 2010 ) .
y. levin , `` internal thermal noise in the ligo test masses : a direct approach , '' phys .
rev .
d * 57 , * 659663 ( 1998 ) .
g. m. harry , a. m gretarsson , p. r. saulson , s. e. kittelberger , s. d. penn , w. j. startin , s , rowan , m. m. fejer , d. r. m. crooks , g. cagnoli , j. hough , and n nakagawa , `` thermal noise in interferometric gravitational wave detectors due to dielectric optical coatings , '' class .
quantum grav .
* 19 , * 897917 ( 2002 ) .
r. nawrodt , s. rowan , j. hough , m. punturo , f. ricci , and j .- y .
vinet `` challenges in thermal noise for 3rd generation of gravitational wave detectors , '' gen relativ gravit * 43 , * 593622 ( 2011 ) . a. bunkowski , o. burmeister , d. friedrich , k. danzmann and r. schnabel , `` high reflectivity grating waveguide coatings for 1064 nm , '' class . quantum grav .
* 23 , * 72977303 ( 2006 ) .
f. brckner , t. clausnitzer , o. burmeister , d. friedrich , e .- b .
kley , k. danzmann , a. tnnermann , and r. schnabel , `` monolithic dielectric surfaces as new low - loss light - matter interfaces , '' opt .
lett .
* 33 , * 264266 ( 2008 ) .
f. brckner , d. friedrich , t. clausnitzer , o. burmeister , m. britzger , e .- b .
kley , k. danzmann , a. tnnermann , and r. schnabel , `` demonstration of a cavity coupler based on a resonant waveguide grating , '' optics express * 17 , * 163169 ( 2009 ) .
f. brckner , d. friedrich , t. clausnitzer , m. britzger , o. burmeister , k. danzmann , e .- b .
kley , a. tnnermann , and r. schnabel , `` realization of a monolithic high - reflectivity cavity mirror from a single silicon crystal , '' phys .
rev .
lett .
* 104 , * 163903 ( 2010 ) .
m. p. edgar , b. w. barr , j. nelson , m. v. plissi , k. a. strain , o. burmeister , m. britzger , k. danzmann , r. schnabel , t. clausnitzer , f. brckner , e .- b .
kley , and a. tnnermann , `` experimental demonstration of a suspended diffractively coupled optical cavity , '' opt . lett . *
34 * 3184 - 3186 ( 2009 ) .
b. w. barr , s. h. huttner , j. r. taylor , b. sorazu , m. v. plissi , and k. a. strain , `` optical modulation techniques for length sensing and control of optical cavities , '' appl . opt . * 46 * 7739 - 7745 ( 2007 ) .
s. h. huttner , b. w. barr , m. v. plissi , j. r. taylor , b. sorazu , and k. a. strain , `` novel sensing and control schemes for a three - mirror coupled cavity , '' class . and quantum grav . , * 24 * , 3825 ( 2007 ) .
a. hessel and a. a. oliner , `` a new theory of wood s anomalies on optical gratings , '' appl . opt .
* 4 , * 12751297 ( 1965 ) .
l. mashev and e. popov , `` diffraction efficiency anomalies of multicoated dielectric gratings , '' opt
. commun .
* 51 , * 131136 ( 1984 ) .
l. mashev and e. popov , `` zero order anomaly of dielectric coated gratings , '' opt .
commun .
* 55 , * 377380 ( 1985 ) .
g. a. golubenko , a. s. svakhin , v. a. sychugov , and a. v. tishchenko,``total reflection of light from a corrugated surface of a dielectric waveguide , '' sov .
j. quantum electron * 15 , * 886887 ( 1985 ) .
r. magnusson and s. s. wang , `` new principle for optical filters , '' appl .
. lett .
* 61 , * 10221024 ( 1992 ) . c. f. r. mateus , m. c. y. huang , l. chen , c. j. chang - hasnain , and y. suzuki , `` broad - band mirror ( 1.12 - 1.62)@xmath7 m using a subwavelength grating ) , '' ieee photonics technology letters * 16 , * 16761678 ( 2004 ) .
f. brcker , d. friedrich , m. britzger , t. clausnitzer , o. burmeister , e .- b .
kley , k. danzmann , a. tnnermann , and r. schnabel `` encapsulated subwavelength grating as a quasi - monolithic resonant reflector , '' optics express * 17 , * 2433424341 ( 2009 ) .
a. sharon , d. rosenblatt , and a. a. friesem , `` resonant grating - waveguide structures for visible and near - infrared radiation , '' j. opt .
soc .
am . a * 14 , * 29852993 ( 1997 ) . m. g. moharam and t. k. gaylord , `` rigorous coupled - wave analysis of planar - grating diffraction , '' j. opt
. soc
. am .
* 71 , * 811818 ( 1981 ) .
m. v. plissi , c. i. torrie , m. e. husman , n. a. robertson , k. a. strain , h. ward , h. lck , and j. hough , `` geo 600 triple pendulum suspension system : seismic isolation and control , '' rev .
sci .
instrum .
* 71 , * 25392545 ( 2000 ) .
r. w. p. drever , j. l. hall , f. v. kowalski , j. hough , g. m. ford , a. j. munley , and h. ward , `` laser phase and frequency stabilization using an optical resonator , '' appl .
phys .
b * 31 , * 97105 ( 1983 ) . a. e. dangor and s. j. fielding , `` the response of the fabry - perot interferometer to rapid changes in optical length , '' j. phys .
d : appl .
phys .
* 3 , * 413421 ( 1970 ) .
b. w. barr et al . , in preparation |
consider a finite volume containing interacting particles and that is coupled at opposite boundaries to particle reservoirs held at _ different _ chemical potentials , as shown in figure [ fig : chempotgrad ] .
in this situation one anticipates a net flux of particles from the reservoir with the greater chemical potential through the system and into the opposite reservoir . after some initial transients we expect a state to arise
in which there is a nonzero mean flux that is constant over space and time .
this flux reveals the system to be in a _
nonequilibrium steady state_. more generally , we have in mind any physical system all of whose observables do not change with time , but nevertheless exhibits an irreversible exchange of heat , particles , volume or some other physical quantity with its environment . a channel held out of equilibrium through the application of a chemical potential gradient .
it is anticipated that a nonequilibrium steady state will be reached in which particles flow from one reservoir to another at a constant rate . ]
when heat , particle or volume exchanges are reversible , the system is at equilibrium with its environment and , as is well established , its microstates have a gibbs - boltzmann distribution .
this knowledge allows one then to predict the macroscopic physics of a many - body system given a microscopic model couched in terms of energetic interactions .
furthermore , these gibbs - boltzmann statistics emerge from the application of a very simple principle , namely that of equal _ a priori _ probabilities for the system combined with its environment . by contrast ,
when a system is out of equilibrium , very little is known about the statistics of the microstates , not even in the steady state .
this is despite a wide range of approaches aimed at addressing this deficiency in our understanding of nonequilibrium physics .
these approaches fall roughly into two broad categories .
first , one has macroscopic theories , such as onsager and machlup s pioneering work on near - equilibrium fluctuations @xcite . in this regime , it is assumed that the thermodynamic forces that restore an equilibrium are linear , and it is possible to determine the probability of witnessing certain fluctuations in the equilibrium steady state , the most probable trajectory given by a principle of minimum energy dissipation . in this work ,
we are interested in steady states that are far from equilibrium , i.e. , where the system is driven beyond the linear response regime .
recently there has been some success in extending the onsager - machlup theory of equilibrium fluctuations to such nonequilibrium steady states , as long as one has been able to derive macroscopic ( hydrodynamic ) equations of motion for the system @xcite .
furthermore , over the past few years fluctuation theorems have been derived under a wide range of conditions . typically , these relate the probabilities of entropy changes of equal magnitude but opposite sign occurring in a system driven arbitrarily far from equilibrium ( see @xcite for a brief introductory overview ) .
although some connections between these various topics have been established ( see , e.g. , @xcite ) , it is fair to say that a complete coherent picture of the macroscopic properties of nonequilibrium steady states is still lacking .
somewhere between this macroscopic approach and a truly microscopic approach ( i.e. , one that would follow directly from some microscopic equations of motion ) lie a range of mesoscopic models that are intended to capture the essential features of nonequilibrium dynamics
irreversibility , currents , dissipation of heat and so on but are nevertheless simple enough that analytical treatment is possible . in particular , it is possible to explore the macroscopic consequences of the underlying dynamics , in the process identifying both similarities with equilibrium states of matter and the novel features peculiar to nonequilibrium systems .
it is this approach that is the focus of this review article .
specifically , we will discuss a set of models that have a steady - state distribution of microstates that can be expressed mathematically in the form of a matrix product .
we will explain in detail how these expressions can be used to calculate macroscopic steady - state properties exactly .
these will include particle currents , density profiles , correlation functions , the distribution of macroscopic fluctuations and so on .
these calculations will reveal phenomena that occur purely because of the far - from - equilibrium conditions : for example , boundary induced phase transitions and shock fronts . meanwhile ,
we will also find conceptual connections with equilibrium statistical physics : in some of the models we discuss , for example , it is possible to construct partition functions and free - energy - like quantities that are meaningful both for equilibrium and nonequilibrium systems .
we begin by explaining this modelling approach in more detail . in figure
[ fig : microtubule ] we have sketched some molecular motors ( kinesins ) that are attached to a microtubule .
this is essentially a track along which the motors can walk by using packets of energy carried by atp molecules present in the surroundings .
clearly , at the most microscopic level , there are a number chemical processes at play that combine to give rise to the motor s progress along the track . at a more
coarse - grained , mesoscopic level , we can simply observe that the motor takes steps of a well - defined size ( approximately 8 nm for a kinesin @xcite ) at a time , and thus model the system as a set of particles that hop _ stochastically _ between sites of a lattice .
this stochastic prescription is in part intended to reflect the fact that the internal degrees of freedom that govern when a particle hops are not explicitly included in the model .
kinesins attached to a microtubule that move in a preferred direction by extracting chemical energy from the environment . at a mesoscopic level , this can be modelled as a stochastic process in which particles hop along a one - dimensional lattice as shown . ]
the simplest such stochastic model for the particle hops is a poisson process that occurs at some prescribed rate .
let @xmath0 and @xmath1 be two configurations of the lattice that differ by a single particle hop .
we then define @xmath2 as the rate at which this hop occurs , such that in an infinitesimal time interval @xmath3 the probability that that hop takes place is @xmath4 .
the rate of change of the probability @xmath5 for the system to be in configuration @xmath0 at time @xmath6 is then the solution of the master equation @xmath7 subject to some initial condition @xmath8 .
the first term on the right - hand side of this equation gives contributions from all possible hops ( transitions ) into the configuration @xmath0 from other configurations @xmath1 ; the second term gives contributions from hops out of @xmath0 into other configuration @xmath1 .
we are concerned here with steady states , where these gain and loss terms exactly balance causing the time derivative to vanish .
such solutions of the master equation we denote @xmath9 . in this work ,
nearly all the models we discuss have the property that a single steady state is reached from any initial condition : i.e. , they are ergodic @xcite . given then that the distribution @xmath9 is unique for these models , we concentrate almost exclusively on how it is obtained analytically , and how expectation values of observables are calculated
. the remaining models that are not ergodic over the full space of configurations turn out to have a unique steady - state distribution over some subspace of configurations .
we will see that such a distributions of this type can also be handled using the methods described in this work .
solutions of the master equation ( [ int : me ] ) can assume a number of structures .
some general classes of steady state not intended to be mutually exclusive can be summarised as follows : [ [ equilibrium - gibbs - boltzmann - steady - state ] ] equilibrium ( gibbs - boltzmann ) steady state + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + if the internal energy of a microstate @xmath0 is @xmath10 and the system as a whole is in equilibrium with a heat bath with inverse temperature @xmath11 , the microstates have a gibbs - boltzmann distribution @xmath12 .
the dynamics is reversible in that the probability of witnessing any particular trajectory through phase space is equal to that of its time reversal .
this then implies that stochastic dynamics expressed in terms of transition rates @xmath2 must satisfy the detailed balance condition @xcite @xmath13 where here @xmath9 is the gibbs - boltzmann distribution .
a simple consequence of ( [ int : db ] ) is that there are no fluxes in the steady state .
hence any system that does exhibit currents in the steady state , i.e. , nonequilibrium systems , generally do not have a stationary distribution that satisfies detailed balance . [
[ factorised - steady - state ] ] factorised steady state + + + + + + + + + + + + + + + + + + + + + + + sometimes , whether in or out of equilibrium , one has a factorised steady state .
typically one has in mind a lattice ( or graph ) with @xmath14 sites with configurations specified by occupancy variables @xmath15 ( i.e. , site @xmath16 contains @xmath15 particles ) . a factorised steady state
then takes the form @xmath17 .
for such a structure to arise , certain constraints on the transition rates @xmath2 must be satisfied .
for example , in an equilibrium system , one would need the energy of the system to be the sum of single - site energies .
out of equilibrium , one finds factorised steady states in zero - range processes , in which particles hop at a rate that depends only on the occupation number at the departure site .
moreover , necessary and sufficient conditions for factorisation in a broad class of models which includes the zrp have been established @xcite .
zero - range processes and various generalisations have been reviewed recently in a companion paper @xcite and so we do not discuss them further here .
[ [ matrix - product - steady - state ] ] matrix product steady state + + + + + + + + + + + + + + + + + + + + + + + + + + + a matrix product steady state is an extension to a factorised steady state that is of particular utility for one - dimensional models .
the rough idea is to replace the scalar factors @xmath18 with matrices @xmath19 , the steady state probability then being given by an element of the resulting matrix product @xmath20 . since the matrices @xmath21 for different occupancy numbers @xmath22 need not commute , one opens up the possibility for correlations between the occupancy of different sites ( above those that emerge from global constraints , such as a fixed particle number , that one sees in factorised states ) .
it turns out that quite a number of nonequilibrium models have a matrix product steady state , and we will encounter them all in the course of this review . [ [ most - general - steady - state ] ] most general steady state + + + + + + + + + + + + + + + + + + + + + + + + + in principle , the ( assumed unique ) stationary solution of the master equation ( [ int : me ] ) can always be found if the number of configurations is finite , since ( [ int : me ] ) is a system of linear equations in the probabilities @xmath9 .
one way to express this solution is in terms of _ statistical weights _
@xmath23 , each of which is given by the determinant of the matrix of transition rates obtained by removing the row and column corresponding to the configuration @xmath0 ( see e.g. @xcite ) .
to arrive at the probability distribution @xmath9 one requires the _ normalisation _ @xmath24 , so that then @xmath25 .
this normalisation has some interesting properties : it is uniquely defined for any ergodic markov process with a finite number of configurations ; it can be shown to be equal to the product of all nonzero eigenvalues of the matrix of transition rates ; and it is a partition function of a set of trees on the graph of transitions between different microscopic configurations . in this latter case ,
the densities of particular edges in the ensemble of trees are controlled by tuning the transition rates corresponding to those edges , so in this interpretation these transition rates are equivalent to equilibrium fugacities .
although this interpretation of the normalisation is rather abstract , concrete connections between transition rates in certain nonequilibrium models that have a matrix product steady state and fugacities in an equilibrium ensemble have been made , and these we shall discuss later in this review .
we also remark that the equilibrium theory of partition function zeros to characterise phase transitions also carries over to the normalisation as just defined , at least for those models that have been tested @xcite .
we have three principal aims in this work .
first , we wish to illustrate the insights into nonequilibrium statistical mechanics that have been gained from exactly solvable models , particularly those that have a steady state of the matrix product form .
secondly , we seek to provide a self - contained pedagogical account of the various analytical methods and calculational tools that can be used to go from the matrix product expressions to predictions for the macroscopic physics .
finally , we wish to present a thorough review of the progress that has been made using the matrix product approach over the last few years .
as far as we are aware , the matrix product approach has not been the focus of a review for nearly a decade @xcite , and we feel that it is high time that the significant developments that have occurred in the meantime should be collected together in one place . in order to prevent this review from becoming unmanageably long , we focus purely on static physical properties exhibited in the steady states of models solvable by the matrix product method .
this means we have unfortunately had to omit discussion of some very interesting topics , for example how some dynamical properties , such as fluctuation phenomena , have been elucidated through the use of the bethe ansatz @xcite , determinental solutions and their connection to random matrix theory @xcite and dynamical matrix products@xcite .
some of these topics , however , have recently been reviewed elsewhere @xcite .
we also direct the reader to established reviews , such as @xcite , for any other background that we have been forced to leave out here . to realise the aims stated above , we provide in the next section a general account of
the physics one expects to see in nonequilibrium dynamical systems , and outline the essential ideas underlying the matrix product approach .
thereafter , we go into the details of how the simplest models are solved , and show a number of contrasting ( but equivalent ) approaches that have found application to more complex problems .
these cases then form the material of the remainder of the review , which we round off by posing some open problems for future research .
the asymmetric simple exclusion process ( asep ) is a very simple model of a driven system in one dimension that has biased diffusion of hard - core particles .
we shall discuss two versions : the periodic system and the open system .
the latter is coupled to particle reservoirs at either end so that , as described in the previous section , there is a steady state with a constant particle flux .
the open system has a long history , having first appeared in the literature to our knowledge as a model of biopolymerisation @xcite and transport across membranes @xcite . in the mathematical literature
meanwhile diffusion with collisions between particles was first studied by harris @xcite and the terminology simple exclusion was first defined by spitzer @xcite . over the years , applications to other transport processes have appeared , e.g. , as a general model for traffic flow @xcite and various other theoretical and experimental studies of biophysical transport @xcite .
whilst interesting and important , these many applications are not our primary concern here . also we shall not do justice to the many rigorous mathematical results which have been summarised in the books by liggett@xcite . rather , our interest in the asep lies in its having acquired the status of a fundamental model of nonequilibrium statistical physics in its own right in much the same way that the ising model has become a paradigm for equilibrium critical phenomena .
in particular , the asep and related models show despite their superficial simplicity a range of nontrivial macroscopic phenomena , such as phase transitions , spontaneous symmetry breaking , shock fronts , condensation and jamming . in common with the vast majority of the models we discuss in this review , the asep is defined as a stochastic process taking place in continuous time on a discrete one - dimensional lattice of @xmath14 sites .
hopping is totally asymmetric : a particle sitting to the left of an occupied site hops forwards one site ( as in figure [ fig : microtubule ] ) , each hop being a poisson process with rate @xmath26 .
that is , in an infinitesimal time interval @xmath3 , there is a probability @xmath27 that one of the particles that can hop , does hop , the identity of that particle being chosen at random . in simulation terms
, this type of dynamics corresponds to a _ random - sequential _ updating scheme , in which bonds between particles are chosen at random and then , if there is a particle at the left - hand end of the bond and a vacancy at the right , the particle is moved forwards . in this scheme ,
each update corresponds ( on average ) to @xmath28 units of time ; in the following we take @xmath29 without loss of generality . the definition of the model is completed by specifying the boundary conditions . asymmetric exclusion process with ( a ) periodic and ( b ) open boundary conditions .
the labels indicate the rates at which the various particle moves can occur . ]
the simplest version of the model has a ring of @xmath14 sites as depicted in figure [ fig : asep](a ) : a particle hopping to the right on site @xmath14 lands on site @xmath30 if the receiving site is empty . under these dynamics
, the total particle number @xmath31 is conserved .
furthermore , as we will show in section [ rrproof ] , the steady state is very simple : all configurations ( with the allowed number of particles ) are equally likely .
as there are @xmath32 allowed configurations the probability of any one is @xmath33 .
the steady - state current of particles , @xmath34 , through a bond @xmath35 is given by @xmath26 ( here equal to 1 ) multiplied by the probability that there is a particle at site @xmath16 and site @xmath36 is vacant .
one finds @xmath37 in the thermodynamic limit where @xmath38 at a constant ratio @xmath39 .
one may also ask , starting from a known configuration , how long the system takes to relax to the steady state .
it turns out that this timescale @xmath40 depends on the system size @xmath14 via the relation @xmath41 which defines a _ dynamic exponent _ @xmath42 .
one can show using the bethe ansatz ( not covered here , but see , e.g. , @xcite ) that for the asep on a ring , @xmath43 .
also in the symmetric case where particles hop both to the left and to the right with equal rates , but still with exclusion , one obtains @xmath44 which is the value one would find for a purely diffusive process . to model the interaction of an open system with reservoirs at different densities , poisson processes acting at the boundary sites are added .
specifically , a particle may be inserted onto the leftmost site with rate @xmath45 ( if it is vacant ) , and may leave the system from the rightmost site at rate @xmath11 : see figure [ fig : asep](b ) . here
we will consider only @xmath45 , @xmath46 although there is no problem in considering rates outwith this range
. then we may think of reservoirs at site 0 with density @xmath45 and a reservoir at site @xmath47 with density @xmath48 . to include these moves in the simulation scheme described above
, one would additionally allow the bonds between the system and the reservoirs to be chosen , and perform the particle updates with probability @xmath45 ( entry ) or @xmath11 ( exit ) according to the bond chosen ( assuming @xmath45 and @xmath11 both less than @xmath30 ) .
these moves admit the possibility of a nonzero current in the steady state .
the open system is more interesting in that phase transitions may occur in the steady state . in the following
we shall explore the origins of these phase transitions , both through approximate treatments and an exact solution . in this context
we recognise by direct analogy with equilibrium phase transitions a phase transition as a sudden change in form of macroscopic quantities such as the particle current across a bond , or the density at a site .
it turns out that the particle current plays a central role , analogous to the free energy of an equilibrium system , in determining the nature of the phase transition i.e. the order of the transition is determined by which derivative of the current with respect to some external parameter exhibits a discontinuity .
we remark that other types of updating scheme can also be considered .
for example , in applications to traffic flow or pedestrian dynamics
@xcite , it is more natural to consider parallel dynamics where many particles can hop in concert . the asep remains solvable under certain classes of parallel updating schemes which will be discussed towards the end of this review ( section [ discretetime ] ) .
we begin by reviewing a classical phenomenological theory , first applied to traffic flow @xcite , which serves as a first recourse in our understanding of the phase diagrams of driven diffusive systems .
the idea is to postulate a continuity equation for the local density @xmath49 @xmath50 where @xmath34 is the current of particles .
( note we use here a different nomenclature and notation to @xcite which discusses concentration and flow rather density and current . )
the analysis rests on the key assumption that there is a unique relation , @xmath51 , between the current and local density .
also , it is assumed that there is a maximum current @xmath52 ( the capacity of the road ) at density @xmath53 .
the first assumption implies that ( [ cty ] ) becomes @xmath54 where @xmath55 now , an implicit solution of ( [ cty2 ] ) is of travelling - wave form @xmath56 as can readily be checked by substitution into ( [ cty2 ] ) .
the arbitrary function @xmath57 is determined by the initial density profile @xmath58 .
the interpretation of the solution is that a patch with local density @xmath49 propagates with velocity @xmath59 .
the propagation of such a patch is known as a collective phenomenon known as a _
kinematic wave _ and @xmath60 is a group velocity .
the velocity @xmath61 of a single particle in an environment of density @xmath49 , on the other hand , is defined through @xmath62 one sees that @xmath63 thus , if the single particle velocity is a decreasing function of density , which is what we expect , we find @xmath64 and kinematic waves travel backwards in the frame of a moving particle .
kinematic waves and sharpening of a shock : the figure sketches the evolution of a density profile @xmath65 from an initial profile @xmath66 according to ( [ cty ] ) and illustrates how sharp discontinuities in the profile may develop ( see text for discussion ) . ]
if @xmath60 is a decreasing function of @xmath49 we have the phenomenon of shock formation ( see figure [ fig : kinematic ] ) .
that is , since the patches of an initial density profile travel with different speeds , the low - density regions catch up with higher density regions and discontinuities in the density profile , known as shocks , develop .
strictly , after the formation of a shock the description of the density profile by the first order equation ( [ cty ] ) breaks downs and one has to supplement ( [ cty ] ) with second - order spatial derivatives to describe the shock profile . to deduce the velocity of a shock consider two regions of density @xmath67 to the left and @xmath68 to the right separated by a shock .
mass conservation implies that the velocity ( positive to the right ) of the shock is given by @xmath69 note that if the current density relation has a maximum then one can have a stationary shock , @xmath70 .
a particular form for the current density relationship studied in @xcite and which is relevant to the asep is @xmath71 in this case
@xmath72 at @xmath73 and @xmath34 is symmetric about the maximum . expression ( [ jasep ] ) may be easily understood for the asep by noting that in order for a particle to hop across a bond , the site to the left must be occupied and that to the right empty . for a bond at position @xmath74 ,
the probability of the former is @xmath75 and that of the latter is @xmath76 . assuming these events
are uncorrelated ( which they are not but we shall come to this later ) , we have @xmath77 . in the case ( [ jasep ] ) one can compute the kinematic wave and shock velocities from ( [ vg],[vs ] ) @xmath78 where , as above , @xmath67 and @xmath68 are the densities on either side of the shock .
thus the kinematic wave velocity is negative when @xmath79 and the shock velocity is negative when @xmath80 .
the kinematic wave theory can be used to predict the phase diagram of the open boundary asep , in which distinct steady - state behaviours are demarcated ( figure [ fig : asepmfpd ] ) . the left hand boundary @xmath81 is considered as a reservoir of particle of density @xmath82 and the right hand boundary ( @xmath83 ) as a reservoir of density @xmath84 . associated with these boundary densities
are kinematic waves with velocities @xmath85 in the case where @xmath86 and @xmath87 both kinematic waves propagate into the system .
so , for example , from an initially empty system the kinematic waves of densities @xmath45 and @xmath48 will enter from the left and right of the system and meet somewhere in the middle forming a shock which then moves with velocity @xmath88 if @xmath89 the shock moves to the right hand boundary and density associated with the left hand boundary , @xmath82 , is adopted throughout the bulk of the system . on the other hand
if @xmath90 the shock moves to the left hand boundary and density associated with the right hand boundary , @xmath84 , is adopted throughout the bulk of the system . in the case
@xmath91 , @xmath70 and the shock is stationary . in the stochastic system
the shock , although on average stationary , diffuses around the system and effectively reflects off the boundaries .
the result is that the shock is equally likely to be anywhere in the system . in the case where one of @xmath45 or @xmath92 , the kinematic wave associated with that boundary does not propagate into the system and the kinematic wave which does propagate from the other boundary controls the bulk density .
thus , the boundary with @xmath45 or @xmath93 controls the bulk density finally if both @xmath45 , @xmath94 .
the kinematic waves from both boundaries do not penetrate . to describe this phase one needs to add a diffusive contribution to the current ( [ jasep ] ) i.e. to consider second - order spatial derivative of @xmath49 which we shall do in the next section .
the result is that steady state of the system has density @xmath95 ; the system adopts the maximal current density @xmath73 which is the density associated with kinematic wave velocity zero .
the resulting bulk densities and currents are then as shown in table [ tab : asepmf ] which corresponds to the phase diagram given in figure [ fig : asepmfpd ] .
here we have adopted what has become the standard nomenclature for the three phases the terminology should hopefully be self - explanatory .
.[tab : asepmf ] properties of the asep obtained using the extremal - current principle under a mean - field approximation for the density - current relationship . [ cols="^,^,^,^",options="header " , ] whilst the matrix representation given in ( [ pasep : d1 ] ) and ( [ pasep : e1 ] ) provides a quick route to the normalisation ( [ pasep : zint ] ) , it is less suited to the task of calculating density profiles . here
, progress is made by using a representation that involves the @xmath96-deformed harmonic oscillator algebra , a well - studied mathematical object for which many results are known @xcite .
central to this approach is a pair of creation and annihilation operators @xmath97 and @xmath98 that satisfy a @xmath96-commutation relation @xmath99 and act on basis vectors according to @xmath100 if we take @xmath101 we find from ( [ pasep : qcom ] ) that the reduction relation ( [ pasep : de ] ) is satisfied .
thus this representation looks like @xmath102 in order that ( [ pasep : dv ] ) and ( [ pasep : ew ] ) are also respected ( bearing in mind @xmath103 ) we require @xmath104 where the parameters @xmath105 and @xmath106 are as previously given in ( [ pasep : ab ] ) . thus this representation is the generalisation to @xmath107 of that used in section [ diag ] .
the constant @xmath108 appearing in ( [ pasep : wvnn ] ) is fixed by the convention that @xmath109 .
that is , we require @xmath110 this sum , which converges when @xmath111 , features prominently in the literature on @xmath96-series as it is the @xmath96-analogue of the exponential function @xcite .
when it converges , it can be expressed as an infinite product and @xmath112 .
diagonalisation of the matrix @xmath113 proceeds in much the same way as described above in subsection [ alsalam ] . here ,
the eigenfunctions are a vector of functions @xmath114 with eigenvalue @xmath115 as before .
the polynomials @xmath116 are found to satisfy the three - term recurrence @xmath117 using the same approach as previously .
taking @xmath118 , allows the functions @xmath116 to be identified as @xmath96-hermite polynomials , the properties of which are discussed in @xcite .
of particular importance is the @xmath96-exponential generating function of these polynomials , since this allows computation of the scalar products @xmath119 and @xmath120 that appear during the calculation of the normalisation @xcite .
one finds for the former @xmath121 and the corresponding expression with @xmath122 for the latter . with the knowledge of the orthogonality relation satisfied by the polynomials @xcite
one can write down the identity @xmath123 which is the final part of the jigsaw that is an expression for the normalisation .
one can verify that the same integral ( [ pasep : zint ] ) is obtained using this alternative representation .
the key benefit offered by this approach is that expressions for the difference in density between neighbouring pairs of sites @xmath124 c^{n - i-2 } { | v \rangle}}{z_n}\ ] ] involve the matrix @xmath125 which is diagonal in this representation .
this one sees from ( [ pasep : deaa ] ) and the definitions ( [ pasep : adag ] ) and ( [ pasep : a ] ) .
specifically , @xmath126 { | m \rangle } & = & \frac{{\langle n | } [ \hat{a } \hat{a}^\dagger - \hat{a}^\dagger \hat{a } ] { | m \rangle}}{1-q } \\ &
= & \frac{\left[(1-q^{n+1 } ) - ( 1-q^n)\right]}{1-q } \delta_{n , m } = q^n \delta_{n , m } \;.\end{aligned}\ ] ] to express the density gradient ( [ pasep : delta ] ) in terms of integrals over the @xmath96-hermite polynomials , one inserts two of the identities ( [ pasep : hp2 ] ) , one in front of the combination @xmath127 $ ] and one behind .
this gives rise to a double integral which is analysed for large @xmath14 and @xmath16 in much the same way as discussed above in subsection [ pasymp ] but with additional complexity stemming from the fact that one has the deformation of two contours to contend with .
full phase diagram of the pasep showing additional sub - phases that have different density decay forms . ] the detailed analysis of this double integral was conducted in @xcite . as with the asep ,
one finds additional sub - phases relating to different boundary profiles within the high - density and low - density phases .
the full phase diagram is as shown in figure [ fig : paseppd ] and the density profiles one obtains summarised as follow . *
* maximal - current phase * @xmath128 , @xmath129 .
once again , one has the universal power - law density profile seen both in the asep with open boundaries and on a ring for sufficiently large @xmath45 and @xmath11 .
that is , near the right boundary one has the decay @xmath130 and a similar decay from the left obtained by invoking the particle hole symmetry that was also present in the asep , viz , @xmath131 . *
* low - density phases * @xmath132 . in all of the low - density phases
there is a region of constant density @xmath133 that extends from the left boundary into the bulk .
near the right boundary the decay takes three different forms which define three sub - phases . * * * ld - i * @xmath134 .
here there is an exponential decay with a characteristic length @xmath135 given by @xmath136 \;.\ ] ] this phase is present in the totally symmetric case @xmath137 ( see figure [ fig : aseppd ] ) . * * * ld - ii * @xmath138 .
this phase , which is also present when @xmath137 , has an exponential decay characterised by @xmath139\ ] ] which is additionally modulated by a power - law decay with exponent @xmath140 , as in the totally asymmetric case . * * * ld - iii * @xmath141 .
when @xmath142 , a new phase appears that has a purely exponential decay with a characteristic length @xmath143 \;.\ ] ] * * high - density phases * @xmath144 .
the properties of the high - density phases are obtained from the particle - hole symmetry .
these all have an exponential decay at the left boundary to a constant bulk density that extends to the right boundary .
the foregoing analysis applies to the case @xmath145 , since this is a restriction on the representations of the identity ( [ pasep : asc1 ] ) and ( [ pasep : hp2 ] ) that were required in the analysis .
as we previously noted , the case where @xmath146 is of interest since then the particles in the bulk prefer to hop to the left , even though the boundary conditions enforce a current that flows in the opposite direction .
although the @xmath96-hermite polynomials have an extension to @xmath146 @xcite it is unclear how to handle the resulting integral representation of the normalisation and other matrix product expressions @xcite .
an alternative approach , followed in @xcite , is to note that , if one were actually to evaluate the integral ( [ pasep : zint ] ) by performing , say , a laurent expansion , one would end up with a sum containing a finite number of terms .
this is because this integral is another way to express the result of a direct reordering using the reduction relations ( [ pasep : de])([pasep : ew ] ) , a procedure that is known to terminate .
such an expression is then valid for any @xmath96 , not just @xmath145 as was required to derive the integral representation in the first place . in other words
, we can access the asymptotics for @xmath146 by first fixing @xmath14 at some finite value with @xmath145 , put @xmath146 in the result and only then at the very end take @xmath147 .
it is possible , using a clutch of identities from the theory of @xmath96-series , to find the finite sum implied by ( [ pasep : zint ] ) @xcite .
we do not quote this formula a complicated mixture of ordinary and @xmath96-combinatorial factors here .
rather , we simply state that the leading term in the normalisation for @xmath146 is , for large @xmath14 @xcite , @xmath148 where @xmath149 is a function that is independent of @xmath14 .
note that this result holds for all values of @xmath45 and @xmath11 : there is only one reverse - bias phase . for ( [ pasep : j ] )
we see that in this phase , the current vanishes asymptotically as @xmath150 although the density profile in the reverse bias phase has not been calculated explicitly , one can conjecture its form using a simple argument .
first we expect most of the particles to be in a domain by the left boundary , and most of the vacant sites by the right .
now , if the interface between these particle- and hole - rich regions were closer to the right boundary than the left , a particle at the interface will escape from the right boundary in a shorter time than a hole from the left .
therefore , in this instance the interface will tend to drift to the left . on the other hand , if the interface were closer to the left boundary it would , by the same argument , drift to the right .
therefore , the lattice will typically be half full in the steady state , a feature noticed in computer simulations for small systems .
we finally briefly mention the case of the pasep where particles may enter and leave the system at both boundaries , as shown in figure [ fig : pasep ] .
a representation similar to that used in subsection [ alsalam ] was established in @xcite , where it was also shown that the functions appearing in the eigenvector of the matrix @xmath151 are a complete set of askey - wilson polynomials @xcite .
these are a set of very general polynomials that contain both the al - salam - chihara and @xmath96-hermite polynomials encountered above as special cases .
they have four parameters related to the transition rates @xmath45 , @xmath11 , @xmath152 and @xmath153 as follows @xmath154 \\ b_{\pm } & = & \frac{1}{2\beta } \left [ ( 1-q-\beta+\delta ) \pm \sqrt{(1-q-\beta+\delta)^2 + 4\beta\delta } \right ] \;.\end{aligned}\ ] ] as in subsection [ alsalam ] , the boundary vectors are @xmath155 and @xmath156 , and the polynomials satisfy an orthogonality relation similar to ( [ pasep : ortho ] ) , but with additional factors appearing to accommodate the new parameters @xcite .
these carry through to the expression one eventually obtains for the normalisation @xmath157^n \;.\end{aligned}\ ] ] one recognises a structure similar to ( [ pasep : zint ] ) and the asymptotic analysis via a contour integral detailed in @xcite is basically the same as described in subsection [ pasymp ] . when @xmath145 , one finds the three familiar phases : a maximal - current phase when @xmath158 ; a low - density - phase when @xmath159 and @xmath160 ; and a high - density phase when @xmath161 and @xmath162 .
when @xmath146 the same three phases are manifested as the model is invariant under the exchanges @xmath163 , at least when all four boundary rates are nonzero .
since the physical properties of these phases are very similar to their counterparts in the asep with open boundaries , we do not discuss them further here .
further details , including an analysis of the bulk densities through the introduction of a fugacity as described in subsection ( [ gce ] ) , are given in @xcite .
as further shown in @xcite , the eigenvalue spectrum of the matrix @xmath151 can be used to determine density profiles and correlation functions .
these show once again the universal power - law decay in the maximal current phase , and various exponential decays in the high- and low - density phases .
in this review we have so far mostly concentrated on how the current and density profiles in driven diffusive systems are calculated from the matrix product expressions for the steady state . since one has to hand the entire stationary distribution of microstates , it is in principle possible to calculate any macroscopic quantity of interest .
for example , the existence of a matrix product solution has recently admitted the calculation of the gibbs - shannon entropy in the symmetric exclusion process with open boundaries @xcite . in this section
we take a closer look at a quantity that has received a great deal of recent attention in the context of open driven - diffusive systems , namely free energy functionals which characterise the probability that a deviation from the most likely density profile is seen in a macroscopically large ( but finite ) system . to be more precise , we consider macrostates corresponding to a density profile @xmath75 where @xmath74 lies in the ( one - dimensional ) interval @xmath164 $ ] which contains some large number @xmath14 of lattice sites .
as we shall see explicitly below , each macroscopic profile @xmath75 can be realised in number of ways that grows exponentially with the number of microscopic degrees of freedom @xmath14 .
this exponential growth leads to the key property of statistical mechanical systems that in the thermodynamic limit , @xmath147 , one particular macrostate
let us call this @xmath165is very much more likely to be realised than any other ( except possibly at a phase transition ) .
that is , the probability @xmath166 $ ] of seeing a particular macrostate behaves for large @xmath14 as @xmath167 \sim \exp\left ( - n { \cal f}[\rho ] \right)\ ] ] where @xmath168 $ ] is a functional of the density profile that vanishes when @xmath169 . in an equilibrium system , @xmath168 $ ] is essentially the free energy of the macrostate , relative to the most likely profile .
of course , in a nonequilibrium steady state , the probability of seeing a deviation from the most like profile can be measured , and so by analogy with the equilibrium case , we refer to the object @xmath168 $ ] that results as a free energy functional .
progress in calculating such free energy functionals and related quantities has been summarised in the lecture notes of derrida @xcite . in this subsection
, we will sketch the derivation of the free energy functional for the nonequilibrium symmetric simple exclusion process ( ssep ) with open boundaries that has been achieved through the matrix product solution for the steady - state distribution @xcite .
this version of the model has symmetric hopping of a single species at unit rate in the bulk , entry at the left and right boundaries at rate @xmath45 and @xmath153 respectively , and exit at the left and right boundaries at rate @xmath11 and @xmath152 .
in other words , we have the dynamics illustrated in figure [ fig : pasep ] of section [ pasep ] with the bias @xmath170 . in order to illustrate the procedure by which the free energy functional was first calculated in @xcite from the matrix product expressions
, we shall follow it here for the much simpler case in which the model parameters are chosen so as to give rise to an _ equilibrium _ steady state . to find this set of parameters
, we employ a kolmogorov criterion @xcite .
this states that if , for every loop @xmath171 in configuration space , the equality @xmath172 holds , the steady state of the system is an equilibrium state for which the detailed balance relation @xmath173 also holds for every pair of configurations @xmath0 and @xmath1 between which transitions occur at nonzero rate .
here we find that ( [ free : kol ] ) is automatically satisfied unless , in total , some nonzero number @xmath174 particles exits at one boundary and re - enters at the other to return to the starting configuration going one way round the loop .
then , the ratio of the two products in ( [ free : kol ] ) is @xmath175 .
therefore , to realise an equilibrium steady state one must have @xmath176 . since the bulk hopping occurs at the same rate in both directions , the detailed balance relation ( [ free : db ] ) implies that all configurations that have @xmath31 particles on the lattice are equally likely in the steady state . by considering a pair of configurations that differ by the addition or removal of a single particle at the left boundary
, one finds that if @xmath177 is the steady - state weight of any one of the @xmath31-particle configurations , @xmath178 hence , the probability of seeing a particular configuration @xmath179 ( where , as previously , @xmath180 if site @xmath16 is occupied , and zero otherwise ) is given by the product @xmath181 once the underlying distribution of microstates has been established , the free energy functional is obtained by taking an appropriate combination of thermodynamic and continuum limits . to this end
, we divide the system into a number @xmath22 of boxes , box @xmath182 containing @xmath183 sites and @xmath184 particles .
the @xmath182 pairs @xmath185 can thus be used to specify a macrostate of the system in the limit @xmath147 . to find the probability of this macrostate we note that @xmath186 where @xmath187 is the probability that @xmath14 particles are in a box of size @xmath31 in the steady state .
this is given by @xmath188 \right)\end{aligned}\ ] ] where the second expression , valid for large @xmath14 , has been obtained using stirling s approximation . inserting this into ( [ free : boxes ] ) , and taking the thermodynamic limit , one finds that @xmath189\ ] ] in terms of the intensive box sizes @xmath190 and densities @xmath191 .
the continuum limit is now straightforward : we take @xmath192 and @xmath193 where @xmath194 is some portion of the interval @xmath195 $ ] such that @xmath196 and the density profile @xmath75 becomes a function of @xmath74
. this procedure yields @xmath197 = \int_0 ^ 1 \rmd x \left [ \rho(x ) \ln \frac{\rho(x)}{\bar{\rho } } + ( 1 - \rho(x ) ) \ln \frac{1-\rho(x)}{1-\bar{\rho } } \right]\ ] ] for the free energy functional , where we have introduced the parameter @xmath198 .
it is easily verified that @xmath199 $ ] vanishes if @xmath200 $ ] , indicating that the flat profile is the most likely profile in the thermodynamic limit as one would expect at equilibrium .
we also note that the free energy functional is local in the density field @xmath75 ; that is , one can express ( [ free : eqm ] ) as an integral over a free energy density @xmath201 that depends only on the local properties of the density field @xmath75 at the point @xmath74 ( and nowhere else ) . when the equilibrium constraint on the boundary rates @xmath176 is relaxed , this local property of the free energy density no longer holds .
the reason for this is that when the system is divided into boxes , the joint distribution of box sizes and occupancy does not factorise as in ( [ free : boxes ] ) .
that is , there are long - range correlations in the density profiles that are implied by the matrix - product structure of the steady - state weights .
this makes evaluating the free energy functional by taking the appropriate combination of thermodynamic and continuum limits a much more complicated exercise .
we therefore omit the technical details of the calculation here , referring the reader to @xcite for a full account .
the way in which the free energy functional is obtained in @xcite is to introduce a pair of fugacities for each box , one of which controls its mean size @xmath202 and the other its density @xmath203 .
this procedure is similar to the use of the grand canonical ensemble described in section [ gce ] , albeit with a much larger number of fugacities , and yields the generating function of the joint distribution that extends to the nonequilibrium case the expression ( [ free : boxes ] ) .
the values of the fugacities are then fixed by finding the saddle - point of this generating function .
it turns out that the resulting expression for the free energy functional , @xmath204 = \int_0 ^ 1 \rmd x \left [ \rho(x ) \ln
\frac{\rho(x)}{\sigma(x ) } + ( 1 - \rho(x ) ) \ln \frac{1 - \rho(x)}{1-\sigma(x ) } + \ln \frac{\sigma'(x)}{\sigma(1)-\sigma(0 ) } \right ] \;,\ ] ] involves a companion function @xmath205 which is related to the fugacities described above , and that is fixed by the saddle - point analysis as the function that maximises the functional @xmath206}$ ] for a given profile @xmath75 .
one can verify , by asking for a vanishing functional derivative @xmath207 , that this function must satisfy the nonlinear differential equation @xmath208 additionally , the detailed analysis presented in @xcite shows that @xmath205 must further be a monotonic function , and satisfy the boundary conditions @xmath209 we note that the monotonicity requirement on @xmath205 ensures that the argument of the final logarithm appearing in ( [ free : ssep ] ) is positive , and further that the value of the companion function @xmath205 at some point @xmath74 will in general depend on the _ entire _ profile @xmath75 through the differential equation ( [ free : comp])the free energy density is non - local , as previously claimed .
the companion function @xmath205 that appears in the nonequilibrium free energy functional ( [ free : ssep ] ) is somewhat mysterious at a first encounter .
however , a physical meaning can be ascribed to this function , starting from the observation that the values imposed at the boundaries coincide with the mean densities of the left and right boundary sites in the steady state , and hence also the densities of the particle reservoirs at each end of the system . in the bulk
, one finds that an infinitesimal segment of the system located at a point @xmath210 can be thought of as being in local equilibrium with a reservoir at density @xmath205 . to see this , it is convenient first to generalise the functional ( [ free : ssep ] ) to a density profile @xmath75 defined on an arbitrary interval @xmath211 $ ] in such a way that if the profile is translated and scaled so that it covers a different interval @xmath212 $ ]
, the free energy per unit length is unchanged .
the functional that has this property and coincides with ( [ free : ssep ] ) for the case @xmath213 is @xmath214}([\rho ] ; \sigma(a ) , \sigma(b ) ) = \int_a^b \rmd x \left [ \rho(x ) \ln \frac{\rho(x)}{\sigma(x ) } + ( 1 - \rho(x ) ) \ln \frac{1 - \rho(x)}{1-\sigma(x ) } + \ln \frac{(b - a)\sigma'(x)}{\sigma(b)-\sigma(a ) } \right ] \;.\ ] ] in the limit of an infinitesimal interval , @xmath215 , this expression approaches the corresponding expression for the equilibrium system @xmath216}([\rho ] ; \bar{\rho } ) = \int_a^b \rmd x \left [ \rho(x ) \ln \frac{\rho(x)}{\bar{\rho } } + ( 1 - \rho(x ) ) \ln \frac{1 - \rho(x)}{1-\bar{\rho } } \right]\ ] ] if we identify @xmath217 with @xmath218 , the density of the reservoirs coupled to the equilibrium system .
the observation that an effective local thermal equilibrium applies does not itself allow one to recover the full expression ( [ free : ssepg ] ) for the free energy functional since _ a priori _ one does not know what densities the effective intermediate reservoirs should have .
however , it was noticed in @xcite that one can construct the free energy functional through an additivity principle that involves the modified free energy @xmath219}([\rho ] ; \sigma(a ) , \sigma(b ) ) = { \cal f}_{[a , b]}([\rho ] ; \rho_a,\rho_b ) + ( b - a ) \ln j(\sigma(a ) , \sigma(b))\ ] ] where @xmath220 is the steady - state current through a system of length @xmath221 coupled to a reservoir of density @xmath217 at the left boundary , and of density @xmath222 at the right . for the ssep ,
this current is @xmath223 where we have assumed that @xmath224 .
( since the dynamics are symmetric , the case @xmath225 can be treated by making the replacement @xmath226 . )
the additivity principle given in @xcite relates the free energy functional for a system to that of two subsystems created by inserting a reservoir at a point @xmath227 .
it reads @xmath228}([\rho ] ; \sigma(a ) , \sigma(b ) ) = \max_{\sigma(y ) } \left\ { { \cal h}_{[a , y]}([\rho ] ; \sigma(a ) , \sigma(y ) ) + { \cal h}_{[y , b]}([\rho ] ; \sigma(y ) , \sigma(b ) ) \right\ } \;.\ ] ] that is , @xmath229 is the density one should choose for the reservoir placed at the point @xmath230 so that the combined free energy of the two subsystems is maximised .
one can verify by recursively subdividing the subsystems created by inserting intermediate reservoirs , and by assuming that each infinitesimal segment created in this way is in a local equilibrium with its boundary reservoirs , that the expression ( [ free : ssep ] ) results .
the fact that one is looking for a maximum in ( [ free : add ] ) further implies that the reservoir densities @xmath205 must satisfy the differential equation ( [ free : comp ] ) .
thus the additivity principle ( [ free : add ] ) in tandem with the effective local equilibrium property and the free energy functional given by ( [ free : comp ] ) are equivalent .
it is worth reiterating a few properties exhibited by the free energy functional ( [ free : ssep ] ) noted in @xcite .
first , as is required , the equilibrium version ( [ free : eqm ] ) is recovered in the limit where the right boundary reservoir density @xmath231 approaches that of the left @xmath232 . to see this , we write ( without loss of generality ) for the effective reservoir density in the bulk @xmath233
x + \delta\sigma(x ) \;.\ ] ] if one substitutes this expression into the differential equation ( [ free : comp ] ) that governs @xmath205 , one finds that the function @xmath234 must be proportional to @xmath235 ^ 2 $ ] if @xmath205 is to remain bounded as @xmath236 . in turn
, this implies that @xmath237 in this limit , yielding the desired equilibrium free energy functional , with @xmath218 equal to both boundary reservoir densities .
it is also straightforward to show that the most probable profile coincides with the stationary profile , which for the ssep is simply the linear function @xmath238 x$ ] .
this is achieved by asking for the functional derivative of ( [ free : ssep ] ) with respect to @xmath75 to vanish .
that is , @xmath239}{\delta \rho(x ) } = \ln \left [ \frac{\rho(x)}{\sigma(x ) } \ , \frac{1-\sigma(x)}{1-\rho(x ) } \right ] = 0 \;,\ ] ] and hence we must have @xmath240 .
the differential equation ( [ free : comp ] ) then implies that the second derivative of @xmath205 must vanish , which ( along with the boundary conditions ) results in the optimal @xmath205 and @xmath75 coinciding with the linear stationary profile .
it is easy to see that then ( [ free : ssep ] ) is zero , indicating that this profile appears with probability @xmath30 in the thermodynamic limit ; it can also be shown @xcite that any other macroscopic profile appears with a probability exponentially small in the number of lattice sites .
it is interesting to compare the magnitude of the free energy functional for a given profile @xmath75 with that for an equilibrium system that has the same stationary profile .
this can be realised by coupling a one - dimensional chain of sites along its length to a series of particle reservoirs that have their chemical potentials tuned in such a way that the desired equilibrium density @xmath165 in a small interval @xmath211 $ ] is attained . since the equilibrium free energy density is local , the free energy functional for the whole system is obtained by summing ( [ free : eqmg ] ) over all these small intervals .
this gives @xmath241 = \int_0 ^ 1 \rmd x \left [ \rho(x ) \ln \frac{\rho(x)}{\bar{\rho}(x ) }
+ ( 1 - \rho(x ) ) \ln \frac{1-\rho(x)}{1-\bar{\rho}(x ) } \right ] \;. \label{free : eq}\ ] ] now ( [ free : ssep ] ) reduces to the equilibrium expression ( [ free : eq ] ) when @xmath242 , the equilibrium profile ( which is linear ) .
however , the nonequilibrium free energy is obtained by maximising the expression ( [ free : ssep ] ) with respect to @xmath243 and generally the maximum is not realised by @xmath244 .
therefore @xmath168 \ge { \cal f}^{\rm eq}[\rho]$ ] which means that the probability of witnessing a particular fluctuation away from the optimal profile for a system driven out of equilibrium is suppressed compared to that for an equilibrium system with the same optimal profile .
this is not always the case however : a similar analysis for the asymmetric simple exclusion process ( see below ) shows that fluctuations can also be enhanced relative to the equilibrium state .
a final further application of the free energy functional is to explore the optimal profiles in the nonequilibrium system after imposing additional global constraints .
for example , one can ask for the most likely profile given that the overall density @xmath245 is fixed .
it turns out @xcite that the the most likely profile has an exponential form and , unless the overall imposed density happens to be equal to the equilibrium density ( in which limit the linear profile is recovered ) , the density is discontinuous at the boundaries @xmath246 .
free energy functionals have also been derived for the partially asymmetric exclusion process ( pasep ) in the case where entry and exit of particles occurs at the left and right boundaries respectively , and the bulk bias is to the right ( i.e. , the forward - bias regime , @xmath145 ) @xcite . in principle , one could follow through the procedure outlined above in subsection [ hardway ] , where one starts with the full stationary distribution of microstates and takes the continuum limit .
it turns out to be more straightforward instead to use the microscopic distribution to extend the additivity principle discussed in subsection [ easyway ] .
then , by the effective local equilibrium property , one can construct the full free energy functional as before .
when the bulk dynamics are asymmetric , it turns out that two versions of the additivity principle come into play .
one applies when the left boundary reservoir density @xmath232 is less than that at the right ; the other when the reverse is true . in both cases
, the additivity relation involves the modified free energy defined by equation ( [ free : mod ] ) , but where now the current is that found ( for example ) by an application of the extremal current principle discussed in section [ asep ] .
that is , @xmath247 } \rho(1-\rho ) & \sigma \le \sigma ' \\
\displaystyle \max_{\rho \in [ \sigma ' , \sigma ] } \rho(1-\rho ) & \sigma \ge \sigma ' \end{array } \right . \;.\ ] ] when @xmath248 , the additivity formula is the same as for the symmetric case , equation ( [ free : add ] ) ( see @xcite for details of the calculation )
. using the local equilibrium property , one then finds the free energy functional to be @xmath249}([\rho ] ; \sigma(a ) , \sigma(b ) ) = -(b - a ) \ln j(\sigma(0),\sigma(1 ) ) + { } \nonumber\\ \!\!\ !
\max_{\sigma(x ) } \int_a^b \rmd x \left [ \rho(x ) \ln ( \rho(x)[1-\sigma(x ) ] ) + ( 1 - \rho(x ) ) \ln [ ( 1 - \rho(x))\sigma(x ) ] \right ] \;.\end{aligned}\ ] ] again , the companion function @xmath205 that gives the effective reservoir densities in the bulk must match the actual reservoir densities at the boundaries , and must also be a nonincreasing function .
when @xmath250 , the additivity formula takes on the different form @xmath251}([\rho ] ; \sigma(a ) , \sigma(b ) ) = \min_{\rho_y \in \{\sigma(a ) , \sigma(b)\ } } \left\ { { \cal h}_{[a , y]}([\rho ] ; \sigma(a ) , \sigma(y ) ) + { \cal h}_{[y , b]}([\rho ] ; \sigma(y ) , \sigma(b ) ) \right\ } \;,\ ] ] where again we refer the reader to @xcite for the details of the calculation .
the fact that the intermediate reservoir inserted at the point @xmath230 always takes on a density that is equal to that of one of the boundary reservoirs implies that effective reservoir density is constant on one side of the point @xmath230 .
further subdivision on that side then has no effect .
on the other side the situation is the same as for the previous subdivision : on one side of the division , the density will be constant and equal to that of the appropriate boundary reservoir , whereas on the other further subdivisions will be required .
the upshot of this is that the reservoir density function @xmath205 will be a step function , taking the value @xmath252 for @xmath253 and @xmath254 for @xmath255 for some value @xmath230 .
this value will be such that the overall free energy is minimised , that is @xmath256}([\rho ] ; \sigma(a ) , \sigma(b ) ) = -(b - a ) \ln j(\sigma(a ) , \sigma(b ) ) + { } \nonumber\\ \!\!\ !
\min_{a\le y\le b } \bigg\{\int_a^y \rmd x \left [ \rho(x ) \ln ( \rho(x)[1-\sigma(a ) ] ) + ( 1 - \rho(x ) ) \ln [ ( 1 - \rho(x))\sigma(a ) ] \right ] + { } \nonumber\\ \hspace{3em } \int_y^b \rmd x \left [ \rho(x ) \ln ( \rho(x)[1-\sigma(b ) ] ) + ( 1 - \rho(x ) ) \ln [ ( 1 - \rho(x))\sigma(b ) ] \right ] \bigg\ } \;.\end{aligned}\ ] ] note that despite considerable cosmetic differences between this formula and ( [ free : asepd ] ) , the two expressions are in fact very similar : in both cases , one needs to find the set of reservoir densities @xmath205 that leads to an extremum of the same joint functional of @xmath75 and @xmath205 .
the fact that in one case , the additivity principle involves a maximum and in another a minimum has its origins in the nature of the saddle - point which provides the relevant asymptotics @xcite .
we also note that the bias parameter @xmath96 does not enter explicitly into these equations , only implicitly through the relationship between the boundary reservoir densities @xmath217 and @xmath222 , and the microscopic transition rates @xmath257 and @xmath96 . specifically @xmath258 and so @xmath96 affects phase boundaries in the @xmath45-@xmath11 plane , as was seen in section [ pasep ] .
the @xmath96-independence of the free energy functionals does not contradict the rich structure seen in the density profiles in section [ pasep ] , since the deviations from the bulk density at the boundaries vanish under rescaling in the @xmath147 limit .
this lack of @xmath96-dependence also implies that ( [ free : ssep ] ) is not obtained as @xmath259 : in other words , the limits @xmath147 and @xmath259 do not commute ( as we have already seen ) .
the free energy functional in the weakly asymmetric limit @xmath260 has been computed explicitly using the matrix product approach and on this scale a @xmath96 ( or @xmath261 ) dependence is apparent @xcite .
closer scrutiny of these free energy functionals for the pasep @xcite shows similar properties to those seen for the ssep above in subsection [ ssepprops ] .
for example , one finds that the stationary profile ( here , a constant profile after rescaling in the thermodynamical limit ) is also the most likely profile , except along the boundary between the high- and low - density phases along which as has been discussed before there is a superposition of shocks with the shock location distributed uniformly across the system .
then , any one of these shock profiles is found to minimise the free energy functional .
again , one can compare the relative size of a fluctuation away from the most likely profile in the nonequilibrium system , and an equilibrium system coupled to a reservoir with a spatially - varying chemical potential .
in the case where @xmath262 , it is found that ( as for the ssep ) , such fluctuations are suppressed ; however when @xmath263 the opposition between the boundary densities and the bulk bias results in these fluctuations being enhanced .
a final interesting feature of the pasep is that density fluctuations in the maximal current phase are non - gaussian .
evidence for this is provided by the existence of a discontinuity at a density of @xmath264 in the probability of witnessing a density @xmath49 in a box located somewhere in the bulk of the system . that the distribution is indeed non - gaussian is confirmed by explicit calculations for the totally asymmetric case ( @xmath137 ) @xcite that exploit the relationship to equilibrium surface models outlined in subsection [ surf ] . in this section
we have seen that once one has found a function that is additive when two subsystems are connected together via a reservoir with an appropriate density , the distribution of density profile macrostates can be constructed given the existence of a local equilibrium property .
this is an appealing approach , and one hopes that it can be applied to a much wider range of systems than exclusion processes .
the difficulty is , however , that one does not know _ a priori _ what form the additivity principle should take : here , we have relied on the complete knowledge of the underlying distribution of microstates to construct it .
nevertheless , it is worth remarking that the free energy functional for the symmetric exclusion process ( [ free : ssep ] ) has been obtained in a purely macroscopic formalism @xcite . in this approach
it is assumed that , in the combined thermodynamic and continuum limit , the macroscopic density profile evolves deterministically as a consequence of the law of large numbers .
then , it can be shown that the probability of witnessing a deviation @xmath75 from the most likely profile @xmath165 in the steady state is given by a functional of the most likely trajectory through phase space that begins at time @xmath265 at stationarity @xmath165 and is constrained to reach @xmath75 at @xmath266 @xcite .
so far we have seen exact matrix product solutions for the steady state of two classes of model systems with nonequilibrium dynamics : the open boundary asep and pasep various aspects of which were discussed in detail in sections [ asep ] , [ open ] , [ pasep ] and [ free]which had a single species of particles hopping on a one - dimensional lattice with open boundaries ; a two - species model with periodic boundary conditions , which we examined in section [ ring ] . in this section
we are going to search for @xmath267-species exclusion models that can be solved using the matrix product approach . for clarity
, we reiterate that in this work the number of species @xmath22 relates to the number of particle species , excluding vacancies .
it is not known how to perform an exhaustive search of _ all _ possible matrix product steady states , but it is possible to search a restricted subset where the matrix products involved can be systematically _ reduced _ using expressions like ( [ int : de])([int : ew ] ) and ( [ ring : de])([ring : da ] ) . to this end , let us recall the proof of the reduction relations that applied for exclusion models on the ring geometry given in subsection [ rrrproof ] .
this proof concerned stationary weights that were given by a trace of matrices @xmath268 where the variable @xmath269 indicates the species of particle occupying site @xmath16 ( @xmath270 denotes a vacancy ) and @xmath271 is the corresponding matrix whose forms is to be determined .
we showed that when @xmath272 is the rate at which particles on sites @xmath16 and @xmath36 exchange places , if one can find auxiliary matrices @xmath273 such that @xmath274 is satisfied , the weights given by ( [ 2s : f ] ) are stationary .
if we are to obtain matrix reduction relations , we require that the auxiliaries @xmath273 are scalars ( rather than matrices ) .
we must also insist that these reduction relations describe an associative algebra , i.e. , that no matter what order the reduction relations are applied , one always ends up with the same sum of irreducible strings .
the various ways in which this can be be achieved were formalised , classified and catalogued by isaev , pyatov and rittenberg @xcite . in this section , we briefly outline the classification scheme for the the case of two particle species , @xmath275 , and show what physical dynamics the various possibilities correspond to .
we then move on to discuss geometries other than the ring . in this section
we will mostly use the notation established above , where vacancies are denoted by @xmath276 and particles by @xmath30 and @xmath267 .
occasionally , it will be helpful to state model dynamics using a natural `` charge representation '' , where particles of species 1 and vacancies are relabelled as positive and negative charges and particles of species 2 as vacancies : @xmath277 note that our notation differs from that used by @xcite .
we will also use a shorthand for the transition rates @xmath278 and , since we are assuming the auxiliary matrices are scalars , we will write them as @xmath279 .
we first restrict ourselves to the case @xmath280 , @xmath281 and @xmath282 , i.e. there are exchanges in at least one direction between particles and vacancies and between the two species of particles .
then , the relations ( [ 2s : m ] ) become @xmath283 it remains to check whether the relations ( [ sqa ] ) are consistent . the approach of isaev , pyatov and rittenberg @xcite
( see also that of karimipour @xcite ) is to generalise the reduction process leading to ( [ int : reord ] ) .
that is , one seeks to use ( [ sqa ] ) to reduce any product of matrices to a sum of irreducible strings .
first we fix the order of irreducible strings as @xmath284 .
then we require that reducing an arbitrary product @xmath285 of the matrices @xmath286 to a sum of irreducible strings @xmath287 is independent of the order in which the reduction is carried out using the elementary rules ( [ sqa ] ) .
reduction rules which satisfy this requirement are referred to by isaev , pyatov and rittenberg as _ pbw - type algebras _ @xcite .
for the two - species case the requirement is that the two possible reductions of the product @xmath288 , illustrated schematically below , give equivalent results .
( 17,4 ) ( 1.5,1.5)@xmath289 ( 4.5,2)(1,1)1 ( 4.5,1.4)(1,-1)1 ( 6,3)@xmath290 ( 6,0)@xmath291 ( 9,3.2)(1,0)1 ( 9,0.2)(1,0)1 ( 10.5,3)@xmath292 ( 10.5,0)@xmath293 ( 13.5,3)(1,-1)1 ( 13.5,0.4)(1,1)1 ( 15,1.5)@xmath294 calculating explicitly using ( [ sqa ] ) yields @xcite @xmath295 which results in the following six conditions on the hopping rates @xmath296 from the requirement that each of the six terms in the above equation vanish : @xmath297 the solutions of these equations were classified in @xcite . here we summarise the various nontrivial solutions which are physically relevant and the corresponding models which have been previously studied in the literature .
the classification hinges on how many of the @xmath298 are zero .
the first class of solutions has none of the @xmath299 .
there are then two possible solutions to ( [ pbw1])([pbw6 ] ) . *
* * solution ai**@xmath300 .
physically , this corresponds to symmetric exclusion with two species which , although labelled distinctly , have identical dynamics .
matrices are easily found by setting @xmath301 which reduces ( [ sqa ] ) to the single species condition @xmath302 . on the ring
this system has a simple steady state where all allowed configurations are equally likely . * * * solution aii**@xmath303 for @xmath304 and @xmath305 for @xmath306 .
this corresponds to a model where the hop rates are totally asymmetric and of the form @xmath307 that is , since @xmath308 , a particle of species 1 is faster than a particle of species 2 and overtakes with rate @xmath309 .
the corresponding matrix algebra is @xmath310 this solution of ( [ pbw1])([pbw6 ] ) was first noted and studied by karimipour @xcite and followed up in @xcite .
as it happens , this model can be generalised to more than two species , and as such will be discussed in its full generality in section [ sec : karimipour ] .
representations are given in appendix [ appreps ] equation ( [ karimrep ] ) .
this second class of solutions has one of @xmath299 .
the first two families of solutions within this class are obtained by taking @xmath311 , and then choosing the hop rates to satisfy ( [ pbw1 ] ) , ( [ pbw3 ] ) and ( [ pbw5 ] ) in two different ways . *
* * solution bi**@xmath312 , @xmath313 , @xmath314 and @xmath315 with @xmath316 .
this corresponds to a model with the hop rates @xmath317 or , in the charge representation , @xmath318 this model is thus an asymmetric generalisation of the second - class particle of section [ ring ] and was first used by @xcite for the special case @xmath319 , @xmath320 .
the corresponding matrix algebra is @xmath321 a representation of this algebra is given in appendix [ appreps ] . * * solution bii * @xmath322 , @xmath323 , @xmath324 , @xmath325 and @xmath326 .
the corresponding model has hop rates @xmath327 or , in the charge representation , @xmath328 this model is another generalisation of the second - class problem first studied with @xmath29 , @xmath137 for a single species two particle @xcite and later generalised in @xcite .
the matrix algebra is @xmath329 this algebra can be mapped onto that used to solve the pasep with open boundaries , ( [ pasep : de])([pasep : ew ] ) , if one takes @xmath330 and @xmath331 , @xmath332 .
thus one can make use of results from section [ pasep ] , along with techniques for models on the ring described in section [ ring ] to analyse various cases of this model @xcite . choosing @xmath333 or @xmath334 to be zero results in the same solution as bii under relabelling particles , and one additional solution when @xmath335 .
this is * * * solution biii**@xmath336 , @xmath337 , @xmath338 , @xmath339 , @xmath340 and @xmath341 .
the corresponding model has hop rates @xmath342 and the matrix algebra is @xmath343 however on the ring this results in a steady state in which all allowed configurations are equally likely : as can be seen from ( [ biiialg ] ) in any periodic string of matrices , all @xmath344 and @xmath345 can be eliminated through the first and third relations
. the third class of solutions is obtained by taking two of the scalar quantities @xmath298 equal to zero . equations ( [ pbw4])([pbw6 ] ) and two of ( [ pbw1])([pbw3 ] ) are then automatically satisfied .
the equation that remains has a structure that is independent of which of the @xmath298 are taken to be nonzero , so we take @xmath346 which leaves us to satisfy ( [ pbw3 ] ) .
this can be done in two ways . * * solution ci * @xmath314 , @xmath315 , @xmath347 , @xmath348 and @xmath349 .
this corresponds to a model in which all six exchanges may take place @xmath350 and that has the matrix algebra @xmath351 unfortunately this algebra is not useful in describing a physical system with periodic boundary conditions . to see this
one can take for example a representation where @xmath352 and @xmath353 which satisfies the first two relations of ( [ cialg ] ) .
however the third relation of ( [ cialg ] ) has the form of a vanishing deformed commutator @xmath354 thus if we commute say a matrix @xmath355 all the way around the ring we will end up with the same matrix product multiplied by a factor @xmath356 where @xmath357 is the number of species 2 particles on the ring and @xmath358 .
thus , only for @xmath359 ( @xmath360 ) can we use this algebra on a periodic system ; alternatively , it can be used for general @xmath361 on a closed segment as we discuss in section [ sec : closeg ] . * * solution cii * equation ( [ pbw3 ] ) can also be satisfied by taking @xmath362 and unrestricted choices for the remaining rates @xmath314 , @xmath315 , @xmath347 , @xmath348 .
this has similar dynamics to the previous model , but with no exchange of species 1 and 2 @xmath363 the matrix algebra is similar to ( [ cialg ] ) but with the third relation absent .
in contrast to the previous case , this algebra can be used to describe the above dynamics on a ring @xcite .
this model and its generalisation to multiple species will be discussed further in section [ sec : multispecies ] . in this case
, we take all @xmath335 which leaves all six hop rates free .
@xmath364 the algebra is a set of deformed commutators @xmath365 explicit forms for @xmath345 , @xmath355 and @xmath344 are discussed in section [ sec : closeg ] . on a periodic system , as with case ci discussed above , we require that a matrix product is left invariant after commuting one of the matrices once around the ring .
the condition for this is that @xmath366^{n_0 } \left[\frac{w_{21}}{w_{12}}\right]^{n_2 } = \left[\frac{w_{02}}{w_{20}}\right]^{n_0 } \left[\frac{w_{12}}{w_{21}}\right]^{n_1 } = \left[\frac{w_{10}}{w_{01}}\right]^{n_1 } \left[\frac{w_{20}}{w_{02}}\right]^{n_2 } = 1 \label{ddbcond}\ ] ] where @xmath15 is the number of particles of species @xmath16 .
these conditions are satisfied for example by the abc model @xcite in the special case where all particle numbers are equal @xmath367 .
the fact that the rhs of ( [ dalg ] ) are all zero implies that to use these algebraic relations , detailed balance must hold in the steady state .
therefore the condition ( [ ddbcond ] ) is actually the condition for detailed balance to hold .
the corresponding energy turns out to be an interesting long - range function @xcite . in the abc model ,
the particles and vacancies are relabelled as @xmath368 and have dynamics @xmath369 where we take @xmath145 .
this corresponds to the choice of rates @xmath370 and @xmath371 .
note that the dynamics are invariant under cyclic permutation of the three particle types , and that @xmath372 particles tend to move to the left of @xmath373 particles , @xmath373 particles tend to move to the left of @xmath151 particles and @xmath151 particles tend to move to the left of @xmath372 particles . in the steady state of this model
this results in a strong phase separation into three domains of @xmath372 , @xmath373 and @xmath151 in the order @xmath374 . in the limit @xmath375 with
@xmath145 fixed , the domains are pure in that far away from the boundary of the domain the probability of finding a particle of a different species ( to that of the domain ) tends to zero .
that is , particles from a neighbouring domain may penetrate only a finite distance . in the case where we have equal numbers of each species condition ( [ ddbcond ] )
is satisfied and one can use the matrix product to calculate the steady state exactly @xcite .
this model also exhibits an interesting phase transition in the weakly asymmetric limit where @xmath96 varies with system size @xmath14 as @xmath376 , thus approaching unity in the thermodynamic limit .
then , according to the value of @xmath11 , the steady state can either order into three domains which are rich in @xmath372 @xmath373 and @xmath151 ( but are not pure domains ) or a disordered phase where the particles are typically in a disordered configuration @xcite . in the previous subsection
, we summarised a classification of all the possible sets of dynamics for which a matrix product state with scalar auxiliaries ( i.e. , one that has a set of reduction relations ) exists with a unique decomposition into irreducible strings of matrices .
we also assessed whether these matrix product states could be used for models on the ring .
we now extend this enquiry to other one - dimensional geometries that can be constructed .
we first consider open boundary conditions , i.e. , those where particles can enter and leave at the boundaries .
the most general dynamics is to have rates @xmath377 at which a particle of species @xmath16 ( or a vacancy if @xmath378 ) is transformed into a particle of species @xmath182 at site @xmath30 and rates @xmath379 at which a particle of species @xmath16 is transformed into a particle of species @xmath182 ( or a vacancy of @xmath380 ) at site @xmath14 .
if we are to have statistical weights of the form ( [ int : mp ] ) , i.e. , @xmath381 we then obtain additional conditions involving the matrices and vectors @xmath382 and @xmath383 .
these may be written as @xmath384 where @xmath385 and @xmath386 . since @xmath387 we require @xmath388 in the open boundary case , typically , in addition to the conditions for the solution classes of section [ ipr ] , one has further constraints on the boundary rates to allow conditions ( [ alphaij],[betaij ] ) to be satisfied
we do not attempt to catalogue all solutions here , rather we point out which of the solution classes of section [ ipr ] may , in principle , be used in an open system and reference some examples .
some general considerations of the open boundary conditions for which one has a matrix product state using the quadratic algebra ( [ sqa ] ) are discussed in the two species cases in @xcite . *
* * class a**solution ai , which corresponds to symmetric exclusion of two differently labelled species with identical bulk dynamics , can be used in the open boundary case where the two species are injected and extracted with different rates , under certain conditions .
solution aii can be combined with certain open boundary conditions that were found by karimipour @xcite .
we shall specify these boundary interactions and discuss this model more thoroughly in section [ sec : multispecies ] . *
* * class b**both solutions bi and bii can be used in models with open boundary conditions .
one case that has been studied is the dynamics implied by solution bii with @xmath137 @xcite .
this is sometimes referred to as the bridge model because one has two particle species with opposite velocities that slow down to exchange places when they meet , rather like cars on a narrow bridge .
unfortunately , the full phase diagram of the bridge model and in particular an interesting broken symmetry region cannot be described by a quadratic algebra ( i.e. , one where the auxiliaries are scalars and reduction relations are implied on the steady - state weights ) .
the bii @xmath389 case has also been studied @xcite . * * * class c**recall that class c comprises solutions where only one of the @xmath298 are nonzero .
hence , it is not possible to satisfy that sum rule ( [ xsum ] ) , and hence the dynamics implied by these solutions can not be solved in an open system using a simple matrix product algebra . * * * class d**algebras within class d can only be used in models with open boundaries under the special conditions of detailed balance , which generally does not hold . on the other hand , as we discuss below , detailed balance _ does _ apply in conservative models with closed boundaries and thus such algebras are relevant there see below .
a closed segment is a one - dimensional lattice of size @xmath14 with _ closed _ boundary conditions .
that is , particles can not enter or leave at the left boundary or right boundaries ( sites 1 and @xmath14 ) . for the moment , let us consider conserving models where particles are neither created nor destroyed in the bulk : the only moves that are allowed are those where particles or vacancies on neighbouring sites exchange places .
it is straightforward to show using a kolmogorov criterion ( [ free : kol ] ) that detailed balance is satisfied in the steady state of any such model .
the reason for this is that in order to create a loop in configuration space , every exchange of a pair must be accompanied by an exchange in the opposite direction ( as long as the dynamics is not totally asymmetric ) .
hence the forward loop contains the same set of exchanges as the reverse , and ( [ free : kol ] ) is satisfied .
some quadratic algebras for closed segments and segments closed at one end and open at the other were discussed in @xcite .
typical configuration in the steady state of the partially asymmetric exclusion process on a closed segment of @xmath14 sites with @xmath145 . ]
the simplest example of this type is the partially asymmetric exclusion process ( pasep ) on the closed segment i.e. just one species of particle @xcite .
the appropriate matrix algebra in this case is @xmath390 taking @xmath145 the configuration with the highest weight will be the configuration with all particles stacked up to the right end of the lattice as shown in figure [ fig : pasep - closed ] .
the weight of any configuration can then be obtained by moving particles to the left from the maximal weight configuration and multiplying the weight by a factor @xmath96 at each move as implied by ( [ qdefcom ] ) .
then the statistical weight of any configuration @xmath391 will be given by @xmath392 and the maximal weight is @xmath393 the steady state weight ( [ closegw ] ) is in fact a boltzmann weight with an energy function @xmath394 and the dynamics respects detailed balance with respect to this weight .
let us now turn to class d models , where the matrix algebra takes the form of three deformed commutators ( [ dalg ] ) .
an explicit form for the operators @xmath355,@xmath344,@xmath395 involves tensor products of matrices which obey deformed commutation relations @xmath396 where @xmath397 so , for example ,
@xmath398 \otimes e(\omega_{21}/\omega_{12})\otimes d = 0 \;.\end{aligned}\ ] ] to obtain statistical weights here , we need to prescribe a suitable contraction operation i.e. appropriate boundary vectors . using the representation of the deformed commutator algebra in appendix [ appreps ] ( [ defcomrep ] )
we see that if there are @xmath399 species 1 particles , @xmath400 species 2 particles and @xmath401 vacancies then @xmath402 will give a nonzero contraction for those configurations with the correct number of each species .
in particular for the reference configuration the weight is @xmath403 in order to investigate the structure of shocks , derrida , lebowitz and speer studied a single second - class particle on an infinite system@xcite .
the idea was to describe the stationary measure as seen from the second - class particle by a matrix product .
the model considered was the partially asymmetric generalisation with dynamics @xmath404 taking the second - class particle to be at the origin and considering a window of @xmath174 sites to the left and @xmath22 sites to the right of the second - class the stationary probabilities ( note , not weights ) are written as @xmath405 a \left [ \prod _ { j=1}^n x_{\tau_j}\right ] { | v \rangle}\ ] ] where @xmath372 is once more the matrix corresponding to a second - class particle . in order for this form to hold for a window of arbitrary size i.e. all values of @xmath22 ,
@xmath406 one requires that @xmath407 where @xmath113 . to ensure that the probabilities are correctly normalised we further require @xmath408 the bulk dynamics falls within the solution class bi , and
so the matrix algebra ( [ bialg ] ) should be used with @xmath409 and @xmath410 .
this particular application has the unusual property that the choice of @xmath411 and @xmath333 has physical consequences .
this is because @xmath411 and @xmath333 can not be scaled out of equations ( [ cw ] ) .
the correct choice is made by insisting that the desired asymptotic densities @xmath412 far to the right of the second - class particle ( @xmath413 ) and @xmath414 far to the left ( @xmath415 ) are obtained .
it turns out that one should take @xmath416 choosing @xmath417 corresponds to a shock profile as soon from the second - class particle with the second - class particle tracking the position of the shock .
the structure of the shock was analysed , in particular the decay to the asymptotic values is exponential .
an interesting result is that the characteristic length of the decay becomes independent of the asymmetry when @xmath418 to obtain a representation of the required matrices and vectors , one can use ( [ birep1])([birep0 ] ) given in appendix [ appreps ] for solution class bi .
then , using ( [ cw ] ) , one can construct @xmath382 , @xmath383 to be eigenvectors @xmath419 with eigenvalue 1 .
the expressions for the boundary vectors are , however , quite complicated in this representation .
an alternative approach , used in @xcite , exploits an infinite - dimensional ( rather than semi - infinite ) representation .
finite dimensional representations along special curves were studied in @xcite . in the foregoing , we have assumed that the bulk dynamics conserve particle numbers . in this section
we show how the matrix algebras which were encountered can be adapted to models that have non - conservative dynamics .
the basic idea is to augment the dynamics with additional processes that move between different sectors ( i.e. , configurations with a particular number of particles ) .
these moves will be constructed in such a way that detailed balance between sectors is satisfied , even though detailed balance does not hold within a sector . as such
, one can realise any prescribed distribution of sectors : we choose weights that correspond to the grand canonical ensemble that was introduced in section [ ring ] as a means to study systems with fixed particle numbers .
we first discuss this method in detail , and then give two concrete examples .
as we have just described , the aim is to construct some dynamics such that the statistical weight of a configuration with @xmath31 particles ( one a designated species ) on an @xmath14-site ring is @xmath420 where @xmath421 are the matrices that give the steady state for some conservative process , and @xmath422 is a fugacity that controls the mean particle number @xmath31 . let us denote transition rates for this conservative process as @xmath423 , and the additional rates that serve to increase or decrease @xmath31 by one as @xmath424 .
the statistical weights must then satisfy the master equation @xmath425 + { } \nonumber\\ \fl\qquad \sum_{{\mathcal{c } } ' } \left [ f({\mathcal{c}}';m-1 ) w_{+}({\mathcal{c } } ' \to { \mathcal{c } } ) - f({\mathcal{c}};m ) w_{-}({\mathcal{c}}\to { \mathcal{c } } ' ) \right ] + { } \nonumber\\ \sum_{{\mathcal{c } } ' } \left [ f({\mathcal{c}}';m+1 ) w_{-}({\mathcal{c } } ' \to { \mathcal{c } } ) - f({\mathcal{c}};m ) w_{+}({\mathcal{c}}\to { \mathcal{c } } ' ) \right ] = 0 \;.\end{aligned}\ ] ] now , the first summation is what appears in the master equation for the conservative process , and since all the weights for fixed @xmath31 are proportional to @xmath426 , this first summation vanishes .
the second and third summations can be made to vanish if we choose @xmath427 where @xmath428 and @xmath429 are the site occupation variables specifying to configurations @xmath0 and @xmath1 that have @xmath31 and @xmath430 particles respectively . as we can see
, this takes the form of a detailed balance relation ( [ free : db ] ) .
the two examples that follow serve to demonstrate .
recall that in section [ ring ] we considered a version of the asep with a second - class particle , i.e. , a model with dynamics @xmath431 this falls within the family of solutions denoted bii above . in the analysis of this model we introduced a fugacity @xmath422 as a trick to expediate calculation of statistical weights , tuning its value to yield the desired mean particle density on the ring in the canonical ensemble .
this ensemble can also be generated using physical dynamics as we have just described .
this is achieved by noting that the algebra let @xmath31 count the number of positive charges ( species @xmath30 particles ) .
we note that the algebra ( [ biialg ] ) implies @xmath432 hence from ( [ dbens ] ) , we see that the grand canonical ensemble is realised physically if we introduce two non - conserving processes @xmath433 where the rates satisfy @xmath434 we remark that @xmath153 and @xmath152 as free parameters that enforce a particular fugacity @xmath422 in the grand canonical ensemble generated by these non - conservative dynamics .
this particular implementation of a model with non - conserving dynamics was used in @xcite to generate a grand canonical ensemble dynamically .
a second non - conserving generalisation of the asep on a ring with a second - class particle is best described using the charge representation . for simplicity , we take the dynamics of bii with @xmath29 and @xmath137 and take @xmath435 .
the case @xmath107 was studied in @xcite .
thus , the conserving dynamics are @xmath436 to these dynamics we seek to add the following non - conserving moves @xmath437 in such a way that the bii matrix algebra ( [ biialg ] ) can be exploited . to achieve this , we require matrices such that ( [ dbens ] ) is satisfied for the non - conserving moves
. that is , @xmath438 where @xmath422 is a fugacity counting the total number of positive and negative charges . to make contact with more familiar problems ,
let us put @xmath439 , @xmath440 and @xmath441 .
we see that the previous equation is satisfied if @xmath442 i.e. , we require @xmath372 to be a projector .
this is achieved using the matrices for the second - class particle problem of section [ ring ] , ( [ ring : de][ring : a ] ) i.e. @xmath443 , where @xmath444 , @xmath445 and @xmath446 .
the statistical weight of a configuration is then given by @xmath447 where @xmath31 is the number of vacancies in the configuration .
in @xcite it was shown that an interesting phase transition arises as @xmath448 is varied .
for @xmath449 there is a tendency to create positive and negative particles and this ultimately leads to a vanishing steady - state density of vacancies . on the other hand for @xmath450
there is a tendency to eliminate positive and negative charges at the left and right boundaries respectively of domains of vacancies .
thus domains of vacancies tend to grow and in the steady state this results in a finite density of vacancies .
the transition is remarkable in that it is a phase transition in a periodic one dimensional system with local dynamics without an absorbing state . here
, we shall use the generating function technique of section [ gfa ] to quickly obtain the asymptotics of the partition function and hence demonstrate the phase transition .
first , notice from ( [ noncon2 ] ) that configurations with no vacancies are dynamically inaccessible .
therefore all configurations have at least one vacancy and we write the partition function as @xmath451 = w { \langle w | } ( d + e + w a)^{n-1}{| v \rangle}\ ] ] ( an unimportant subtlety is that we have ignored the degeneracy factor in placing a given configuration of particles on the periodic lattice ; this factor is bounded from above by @xmath14 and hence does not contribute to the exponential part of @xmath452 which controls the macroscopic physics , as will be seen below . ) we introduce the generating function @xmath453 where we define @xmath454 .
now , we may write @xmath455 where @xmath456 then using the generating function technique and ( [ znu ] ) we obtain @xmath457 the singularities of @xmath458 are a square root singularity at @xmath459 ( coming from @xmath460 ) and a pole at @xmath461 which only exists if @xmath462 . using the results of appendix [ gf ] one quickly obtains the asymptotic behaviour of @xmath452 @xmath463 due to the form ( [ wweight ] ) of the weights , @xmath464 , the density of vacancies , is given by @xmath465 where @xmath466 is the average number of vacancies in the system .
therefore we obtain @xmath467 demonstrating the phase transition in the density of vacancies at @xmath468 .
in the previous section we identified all possible two - species dynamics that have a steady state of matrix product form with a quadratic algebra i.e. obeying relations of the type given by ( [ sqa ] ) .
we deferred the discussion of two families of dynamics , as both can be generalised in a straightforward way to arbitrarily many species of particles .
we now examine these cases in more detail .
further discussions of multi - species quadratic algebras are given in @xcite .
we first consider a variant of the asep on the ring in which each particle has a different forward and reverse hop rate ( and there is no overtaking)@xcite .
this is a generalisation of the two species model , case cii of section [ 2species ] .
there are @xmath31 particles and @xmath469 vacancies on the ring of size @xmath14 .
we give each of the @xmath31 particles an index @xmath470 , and introduce for each @xmath470 a pair of rates @xmath471 and @xmath472 so that the dynamics of particle @xmath470 are @xmath473 the corresponding matrix algebra is a generalisation of class cii of section [ 2species ] to @xmath31 particle species .
it takes the form @xmath474 where we have adopted the notation @xmath475 and @xmath476 .
one possible representation of these matrices is given in appendix [ appreps ] as equation ( [ dmuealg ] ) .
however , it is simplest to perform calculations not with a representation , but directly from the algebra ( [ msalg1 ] ) . on a periodic system
let the steady state weight of a configuration specified by @xmath477 where @xmath478 is the number of empty sites in front of particle @xmath470 , be @xmath479 .
the matrix product form for @xmath480 is @xmath481 \label{msmatf1}\ ] ] using ( [ msalg1 ] ) for @xmath482 in ( [ msmatf1 ] ) gives the relation @xmath483 the procedure is continued using , in sequence , ( [ msalg1 ] ) with @xmath484 to obtain @xmath485 f_n ( n_1,\ldots , n_m ) = \left [ \sum_{i=0}^{m-1 } \frac{1}{p_{m - i } } \prod_{j = m+1-i}^{m } \frac{q_j}{p_j}\right ] f_{n-1 } ( n_1,\ldots , n_m-1)\ ] ] the effect of this manipulation has been to commute a hole initially in front of the particle @xmath31 backwards one full turn around the ring .
the result is that the weight of a configuration of size @xmath14 is expressed as a multiple of the weight of a configuration of size @xmath486 with one hole fewer in front of particle @xmath31 . repeating the commutation procedure for a hole initially in front of particles
@xmath487 implies that the weights are @xmath488 \
\left[1-\prod_{k=1}^{m}\frac{q_{k}}{p_{k } } \right]^{-1 } \ ; \label{diss}\ ] ] and we can take @xmath489 .
an interesting phenomenon that may occur in this system is that of condensation . in that case
the steady state is dominated by configurations where one particle has an extensive value for @xmath478 @xcite .
this is easiest to understand when @xmath490 i.e. we consider only forward hops
. then ( [ diss ] ) reduces to @xmath491 and we may write @xmath492 the weight ( [ bef ] ) clearly has the form of the weight of an ideal bose gas with @xmath493 . here
the ` bosons ' are vacancies and the bose states correspond to particles with the energy of the state determined by the hop rate of the particle .
the equivalent of the density of states for the bose system is the distribution of particle hop rates which we denote @xmath494 .
if there is a minimum hop rate @xmath495 this will correspond to the ground state of the bose system . then
, if for low hop rates the distribution of particle hop rates follows @xmath496 bose condensation will occur for high enough vacancy density or equivalently _ low _ enough particle density . when condensation occurs the slowest particle ( the one with the minimum hop rate @xmath497 ) has an extensive number of vacancies in front of it whereas the rest of the particles will have gaps in front of them comprising some finite number of vacancies .
thus a ` traffic jam ' forms behind the slowest particle . in @xcite
it is shown that the asep on a periodic system may be mapped onto a zero range process and the condensation transition is fully discussed in the context of the zero range process .
recently , further progress has been made in understanding the case where @xmath498 by using extreme values statistics and renormalisation arguments @xcite .
this is the generalisation of the two - species case aii , which had the two particle species hopping with different speeds , and the faster ones overtaking the slower .
the many - species generalisation allows an arbitrary number of particle species labelled @xmath499 .
a particle of species @xmath470 has a forward hop rate or ` velocity ' @xmath500 , and adjacent particles @xmath470 and @xmath501 exchange places with rate @xmath502 if @xmath503 , i.e. , @xmath504 the model provides an elegant generalisation of the open boundary tasep and its algebra .
the generalisation of ( [ aiialg ] ) to the multispecies case is @xmath505 writing @xmath506 , @xmath507 and making the convenient choice @xmath331 , @xmath508 for @xmath509 yields @xmath510 on a periodic lattice the relations ( [ kalg ] ) are the only ones which need be satisfied and this can be done by choosing a scalar representation @xmath511 and @xmath512 where @xmath513 is chosen to be less than the lowest hop rate
. thus on a ring all allowed configurations are equally likely .
more interestingly , the relations ( [ msalg1 ] ) may be used for an open system with suitable boundary conditions . as discussed in section [ sec : extop ] we require @xmath514 which with the above choice of @xmath411 and @xmath515 implies @xmath516 at the left hand boundary a particle of species @xmath470 is injected with rate @xmath517 if the first site is empty
. the choice @xmath518 reduces ( [ alphaij ] ) to @xmath519 at the left hand boundary a particle of species @xmath470 leaves the system with rate @xmath520 .
the choice @xmath521 ensures that explicit representations of the matrices can be constructed ( see appendix [ appreps ] ) and ( [ betaij ] ) becomes @xmath522 relations ( [ kalg ] ) , ( [ kb ] ) and ( [ ka ] ) provide an elegant generalisation of the usual asep relations ( [ int : de])([int : ew ] ) which are recovered when we have just one species of particle with hop rate @xmath523 . in general , an infinite dimensional representation of ( [ kalg ] )
is needed to satisfy the boundary conditions ( [ kb ] ) ; see , for example , ( [ karimrep ] ) in appendix [ appreps ] .
the fully disordered system is realised when each particle that enters has a velocity drawn from some distribution @xmath524 , with support @xmath525 $ ] so that the lowest velocity is @xmath526 , and ( [ avvd ] ) is replaced by @xmath527 thus , when site one is vacant a particle enters with rate @xmath45 and its velocity @xmath528 is assigned according to the distribution @xmath529 giving an effective rate @xmath530 .
therefore we can think of a reservoir of particles with velocities distributed according to @xmath524 each attempting to enter with rate @xmath531 . a particle reaching site @xmath14 exits with rate @xmath532 .
then we can replace @xmath533 in the discrete case by @xmath534 in ( [ aiialgm ] ) . letting
@xmath506 and @xmath535 , @xmath331 and @xmath536 recovers the same algebra ( [ kalg ] ) , ( [ kb ] ) and ( [ ka ] ) with @xmath537 replaced by @xmath538 , e.g. , @xmath539 karimipour @xcite studied , amongst other things , the phase diagram of the open boundary model and found the structure depends on the distribution @xmath524 through the parameter @xmath540 = \frac{1}{v_{\rm min}^2 } - \left\langle \frac{v}{(v - v_{\rm min})^2 } \right\rangle\ ] ] which characterises the form of @xmath524 near the lowest hop rate @xmath526 . if @xmath541 < 0 $ ] the three phases ( low - density , high - density and maximal current ) of the open boundary asep remain although the phase boundaries depend on parameters such as @xmath541 $ ] and @xmath526 . however if @xmath541 \geq 0 $ ] the high - density phase is suppressed and only the low - density and maximal current phases exist .
recall that in section [ algproof ] we set out a general cancellation scheme for matrices that guarantees a stationary solution of a master equation for a process on the ring .
this involved a set of ordinary matrices @xmath421 and their auxiliaries @xmath542 . by restricting to the case of scalar auxiliaries
, we showed , in the case of two particle species in section [ 2species ] , that it was possible to determine all possible sets of dynamics that give rise to a matrix product steady state . in this section
, we shall consider the more complicated case where the auxiliaries are matrices or more complicated operators .
here it is not possible as far as we are aware to perform an exhaustive search for solutions .
however , it has been shown that if one has a one - dimensional lattice with @xmath22 particle species , open boundary conditions and arbitrary nearest - neighbour dynamics in the bulk , there does exist a matrix product solution , involving auxiliaries that in general will not be scalars , obeying the algebraic relations set out in section [ algproof ] . in the next subsection we review this existence proof which is due to krebs and sandow @xcite .
unfortunately , this proof does not lead to any convenient reduction relations , like ( [ int : de])([int : ew ] ) for the asep , that might be in operation , also , the proof is not constructive in that it requires the steady state to already be known in order to construct explicit matrices .
rather , the proof demonstrates that there are no internal inconsistencies in the cancellation mechanism of section [ algproof ] . in the last two subsections we discuss two models of which we are aware that give examples of a matrix product state with matrix auxiliaries .
the existence proof of a matrix product state @xcite ( see also @xcite for the discrete time case ) applies to models with open boundary conditions and arbitrary nearest - neighbour interactions .
let us restate the master equation for a general process in the form introduced in section [ algproof ] .
it reads @xmath543 where the operator @xmath544 applied to the function of @xmath14 state variables @xmath428 generates the gain and loss terms arising from interactions between neighbouring pairs of sites in the bulk , and @xmath545 and @xmath546 do the same at the boundaries . here
we shall consider models that have @xmath22 distinct particle species in addition to vacancies : i.e. , the occupation variables take the values @xmath547 .
the matrix product expressions @xmath548 provide the steady state solution to ( [ formal : me ] ) for some set of matrices @xmath421 and vectors @xmath382 and @xmath383 if one can find a second set of auxiliary matrices @xmath542 such that the relations @xmath549 \cdots { | v \rangle } \\ \label{formal : tilde2 } \hat{h}_l { \langle w | } x_{\tau_1 } \cdots { | v \rangle } = - { \langle w | } \tilde{x}_{\tau_1 } \cdots { | v \rangle } \\
\hat{h}_r { \langle w | } \cdots x_{\tau_n } { | v \rangle } = { \langle w | } \cdots \tilde{x}_{\tau_n } { | v \rangle } \label{formal : tilde3}\end{aligned}\ ] ] hold .
then , one obtains a zero right - hand side of the master equation ( [ formal : me ] ) via a pairwise cancellation of terms coming from ( [ formal : mp])([formal : tilde3 ] ) .
thus , relations ( [ int : h3][int : hr3 ] ) for the asep generalise to @xmath550 relations ( [ ft1][ft2 ] ) give the general form for a cancellation mechanism involving matrix ( or possibly tensor ) auxiliaries .
such a cancellation scheme has been proposed in various contexts @xcite including a generalisation to longer ( but finite ) range interactions @xcite .
some consequences of this scheme have been explored in @xcite . the existence of the matrices @xmath421 and @xmath542 and vectors @xmath382 and @xmath383 appearing in ( [ ft1][ft2 ] ) is proved by constructing them within an explicit representation .
this representation has basis vectors that correspond to configurations of the rightmost @xmath551 sites of the @xmath14-site lattice .
these we denote as @xmath552 .
the vector @xmath383 is then ascribed the role of a vacuum state , and the matrices @xmath421 that of creation operators in such a way that @xmath553 one then defines the vector @xmath382 via the scalar products @xmath554 so that ( [ formal : mp ] ) gives the desired statistical weights . the auxiliary matrices @xmath555 are defined as @xmath556 where the operators @xmath557 and @xmath546 are extended to the full space of all sub - configurations @xmath558 as follows .
first , the bulk operator is defined in terms of the microscopic transition rates @xmath559 as @xmath560 when @xmath561 and @xmath562 , and @xmath563 under the same condition that @xmath561 .
all other elements of these operators are zero .
similarly , at the right boundary we have @xmath564 when @xmath565 , and @xmath566 with these definitions established , the relations ( [ formal : tilde1])([formal : tilde3 ] ) follow after some straightforward manipulations@xcite .
let us take stock of this construction .
we have seen that not only does the existence of a set of vectors and matrices that satisfy ( [ formal : tilde1])([formal : tilde3 ] ) imply a stationary solution of matrix product form of the master equation ( [ formal : me ] ) , but also that such a set of vectors and matrices can always be constructed _ once _
one knows the weights ( [ formal : sp ] ) . as we previously saw in section [ pasep ] for the pasep
, one can find a number of different representations of the matrices @xmath567 and @xmath568 ( which correspond to @xmath355 and @xmath345 in the more general setting ) even when the auxiliary matrices @xmath569 and @xmath570 are the same . in the representation detailed above the auxiliaries are not scalars but rather complicated objects . hence we see that ( i ) there may be many choices of both the matrices @xmath421 and their auxiliaries that correspond to the stationary solution of a single master equation ; ( ii ) matrix reduction relations , like those found for the asep ( [ int : de])([int : ew ] ) , constitute only sufficient conditions on @xmath421 , @xmath382 and @xmath383 , since valid representations where these relations do not hold can be found ( that used in this section provides an example ) ; and ( iii ) the construction of the auxiliary matrices through the generators of the stochastic process in ( [ formal : xtilde ] ) does not necessarily imply the existence of any convenient reduction relations for the @xmath571 matrices that allow , for example , efficient computation of statistical properties . as we saw in section [ ring ] the matrix product state on
the ring involves using a trace operation .
the algebraic relations to be satisfied are given in the general case by ( [ ft1 ] ) .
however , as we saw in section [ 2species ] , sometimes it happens that , although the algebraic relations are consistent , the rotational invariance of the trace operation leads to global constraints on particle numbers .
for example , in the abc model discussed in [ sec : abc ] we found that the matrix product state was consistent only for models in which the numbers of each particle species was the same .
it is essentially this property of models with periodic boundary conditions which has precluded the development of an existence proof for a matrix product state parallel to that given above .
these difficulties are discussed in more detail in @xcite .
we now discuss a specific reaction - diffusion model whose steady state can be represented in matrix product form , but where the auxiliaries @xmath555 are matrices .
this model contains one species of particle and the stochastic processes are diffusion , coagulation and decoagulation .
these last two updates involve the annihilation or creation of a particle adjacent to another particle .
if the bulk rates are chosen in the following way @xcite @xmath572 then for suitable boundary conditions the steady state may be written in matrix product form . originally in @xcite a closed segment was considered .
then , one requires @xmath573 the conditions to be satisfied come from the bulk algebra @xmath574 a four - dimensional representation of the matrices was found @xmath575 d & = & \left ( \begin{array}{cccc } 0 & 0 & 0&0\\ 0&\delta/(\delta + 1 ) & \delta/(\delta+1)&0\\ 0&0 & \delta & 0\\ 0&0&0 & 0 \end{array } \hspace{0.1 in } \right)\ , , \\[0.5ex ] \hat{e } & = & \left ( \begin{array}{cccc } 0 & 0 & q^{-1 } & -(q - q^{-1})^{-1}\\ 0&0 & ( q - q^{-1 } ) & -q\\ 0&0 & \delta ( q - q^{-1 } ) & -\delta q\\ 0&0&0&0 \end{array } \hspace{0.1 in } \right)\ , , \\[0.5ex ] \hat{d } & = & \left ( \begin{array}{cccc } 0 & -\delta q^{-1 } & 0 & 0\\ 0 & -\delta(q - q^{-1 } ) & 0 & 0\\ 0&0 & -\delta ( q - q^{-1 } ) & \delta q\\ 0&0&0 & 0 \end{array } \hspace{0.1 in } \right)\ , , \end{aligned}\ ] ] with boundary vectors @xmath576 where @xmath577 so that @xmath578 .
however there is a subtlety with this model on a closed segment which is that the empty lattice is dynamically inaccessible and itself comprises an inactive steady state .
if we wish to exclude this configuration we should choose @xmath105 and @xmath106 so that @xmath579 vanishes .
this can be accomplished but in doing so renders @xmath105 and @xmath106 @xmath14-dependent .
a first order phase transition occurs at @xmath580 . for @xmath581
the system is in a low - density phase whereas for @xmath582 the system is in a high - density phase .
within the matrix product form of the steady state the transition can be understood as the crossing of eigenvalues of @xmath151 .
this may occur here not because the matrices are of infinite dimension but because they have negative elements ( cf .
the transfer matrices that arise in equilibrium statistical mechanics , as described in section [ ising ] ) .
various other non - conserving models with finite - dimensional matrix product states have been identified .
for example , the bulk dynamics just described has also been studied in the case of an open left boundary , where particles enter with rate @xmath45 and leave with rate @xmath152 and a closed right boundary by jafarpour @xcite .
then , one requires @xmath583 and @xmath584 then if @xmath585 one can find two - dimensional representations of the algebra .
further generalisations of this class of model have been shown to have finite - dimensional matrix product states @xcite . a general systematic approach to determine a necessary condition for the existence of a finite - dimensional matrix product state
has been put forward by hieida and sasamoto @xcite in which more examples are given .
second - class particles were discussed in section [ sscp ] and the matrix product solution to the second - class particle problem was detailed in section [ ring ] .
the matrix algebras for two - species generalisations of the second - class particle arose in the bi and bii solution classes of the classification scheme of section [ 2species ] .
we now consider a natural generalisation of the second - class particle to many classes of particle , labelled @xmath586 with dynamics @xmath587 in this case it turns out one requires more complicated operators and auxiliaries than matrices . for @xmath588 , the case of three particle classes ,
the algebraic relations to be satisfied , ( [ ft1 ] ) , become @xmath589 and @xmath590 the two - class problem , where the last lines of ( [ 3sp ] ) are absent , can be solved ( as we have seen in section [ rrrproof ] ) with scalar choices @xmath591 , @xmath592 and @xmath593 .
it turns out that three - class problem is not a simple generalisation of the two - class case , but instead a much more complication solution is needed @xcite .
matrices which satisfy ( [ 3sp1],[3sp ] ) have been given as @xmath594 the objects @xmath567 and @xmath568 that appear as elements of these infinite - dimensional matrices are themselves infinite dimensional matrices @xmath567 and @xmath568 which satisfy the familiar relation @xmath595
. in other words the @xmath286 are rank four tensors , as indeed are the auxiliaries in this representation .
these were also given by @xcite , and take the form @xmath596 the tensor , or matrix within matrix , form of ( [ mat ] ) is closely related to that of the operators used to determine the steady state current fluctuations in an open system @xcite .
in @xcite , some algebraic relations involving _ only _ the matrices @xmath597 ( i.e. , not the auxiliaries ) were quoted .
these included relations which had triples and quartets of matrices reducing to pairs , and others which transformed triples and pairs to other triples and pairs .
as noted in @xcite , the relations given comprise only a subset of those that would actually be required to perform a complete reduction of any string of matrices to an irreducible form .
hence , a concise statement of the steady state of this system , like that encoded by the reduction relations ( [ int : de])([int : ew ] ) for the asep , is not easy . as such ,
no calculations beyond simple cases@xcite have yet been attempted for the stationary properties of this model .
however recently , generalising a construction by angel @xcite for the two - class problem , ferrari and martin @xcite have shown how to generate the steady state of the multi - class problem .
this construction can then be used to determine operators and auxiliaries to solve the steady state for more than three classes @xcite .
so far we have reviewed interacting particle systems involving continuous time , or equivalently , random sequential dynamics as discussed in section [ asep ] .
however , this is not necessarily the most natural choice of dynamics with which to model some physical systems of interest . for example , in traffic flow and pedestrian modelling @xcite it is desirable that the microscopic constituents are able to move simultaneously this often originates from the existence of a smallest relevant timescale , e.g. reaction times in traffic .
therefore , in these systems , parallel dynamics in which all particles are updated in one discrete timestep , may be used . in this section
we consider the asep under three types of discrete - time dynamics : sublattice , ordered sequential and fully parallel . in all these discrete - time cases
the steady state weights are given by the eigenvector with eigenvalue one of the transfer matrix of the dynamics , which we write in a schematic notation is @xmath598 the transfer matrix @xmath599 applied to the weight @xmath600 generates a sum of terms , each being the weight of a configuration multiplied by the transition _ probability _ in one discrete timestep from that configuration to @xmath601 . in sublattice updating the timestep
is split into two halves : in the first half site 1 , the even bonds @xmath602,@xmath603 and site @xmath604 are updated simultaneously ; in the second half time step the odd bonds @xmath605,@xmath606 are updated simultaneously .
note that we require the total number of sites @xmath607 to be even . in figure
[ fig : sublattice ] we show how this type of updating is applied to partially asymmetric exclusion dynamics with open boundary conditions . in an update of site 1 ,
if site 1 is empty a particle enters with probability @xmath45 . in an update of a bond @xmath608 ,
the two possible transitions are : if site @xmath16 is occupied and site @xmath36 is empty the particle moves forward with probability @xmath26 , or else if site @xmath36 is occupied and site @xmath16 is empty the particle moves backward with probability @xmath96 . in an update of site @xmath14 , if site @xmath14 is occupied the particle leaves the system with probability @xmath11 .
we note that this two - step process avoids the possibility of a conflict occurring ( e.g. , two particles attempting to hop into the same site simultaneously ) . typical configuration in the steady state of the partially asymmetric exclusion process on a closed segment of @xmath14 sites with @xmath145 .
] the asep with sublattice dynamics was first studied in the case of deterministic bulk dynamics ( @xmath29 , @xmath137 ) by schtz @xcite .
a matrix product solution for this case was found by hinrichsen @xcite and a matrix product solution for the general case of stochastic bulk dynamics was found by rajewsky et al @xcite .
we outline this general case here .
one proceeds by constructing the transfer matrix for the whole timestep as a product of operators for each half - timestep @xmath609 where @xmath610 @xmath611 when acting on @xmath612 generate the weights of configurations from which @xmath613 could have been reached by a transition associated with that bond or boundary site , multiplied by the transition probability
. we may write @xmath614 \hat{l } = \left ( \begin{array}{cc } 1-\alpha&0\\ \alpha & 1 \end{array } \right)\qquad \hat{r } = \left ( \begin{array}{cc } 1&\beta\\ 0 & 1-\beta \end{array } \right)\label{lrrep}\end{aligned}\ ] ] where the basis for @xmath599 is @xmath615,@xmath616,@xmath617,@xmath618 and for @xmath619 and @xmath620 is @xmath276,@xmath30 .
so , for example , @xmath621 the matrix product state for sublattice updating takes the form @xmath622 in which , for @xmath16 odd , @xmath271 is a matrix @xmath567 if site @xmath16 is occupied by a particle ( @xmath180 ) or a matrix @xmath568 otherwise ; and for , @xmath16 even , @xmath623 is a matrix @xmath624 if site @xmath16 is occupied by a particle or a matrix @xmath625 otherwise .
note that different matrices , hatted and unhatted , are used for the odd and even sublattices .
the cancellation mechanism is as follows @xmath626 \;=\ ; \hat{x}_{\tau_i } x_{\tau_{i+1 } } \label{slcancel } \\
\langle w| \hat{l } \hat{x}_{\tau_1 } \;=\ ; \langle w| x_{\tau_1}\qquad \hat{r } x_{\tau_{2l } } |v\rangle\;=\ ; \hat{x}_{\tau_{2l } } |v\rangle \ , .
\label{slbccancel}\end{aligned}\ ] ] note that the action of the first half time step transfer matrix @xmath627 is to put hatted matrices at the even sites and unhatted matrices on the odd sites , then the action of @xmath628 is to restore them to their original sites .
thus , the matrix product state ( [ slmp ] ) is indeed an eigenvector with eigenvalue one of the full transfer matrix ( [ fulltm ] ) . using the form ( [ trep],[lrrep ] ) of the operators ,
the mechanism ( [ slcancel ] , [ slbccancel ] ) implies the following algebraic relations @xmath629 & = & [ d,\hat{d } ] \;=\ ; 0 \nonumber \\ \hat{e}d & = & ( 1-q)e\hat{d } + p d \hat{e } \\
\hat{d}e & = & ( 1-p ) d \hat{e } + q e \hat{d } \nonumber\end{aligned}\ ] ] and the boundary conditions @xmath630 \langle w| ( \alpha \hat{e}+\hat{d } ) \;=\ ; \langle w| d \end{array } \qquad
\begin{array}{c } ( 1-\beta)d|v\rangle \;=\ ; \hat{d}|v\rangle \\[1 mm ] ( e+\beta d)|v\rangle \;=\ ; \hat{e}|v\rangle \end{array}\,.\ ] ] remarkably , the same matrices and vectors @xmath631 as in the random sequential case can be used to solve these relations . to see this we make the ansatz @xcite @xmath632 where @xmath261 is a scalar to be fixed , and we find that ( [ spbulk ] , [ spbc ] ) reduce to @xmath633 \label{rs1}\\ d|v\rangle \;=\ ; \frac{\lambda}{\beta}|v\rangle \quad ; \quad \langle w |e \
; = \ ; \langle w |\frac{\lambda ( 1-\alpha)}{\alpha}\;. \label{rs2}\end{aligned}\ ] ] finally we may set @xmath634 and define @xmath635 which then satisfy @xmath636 thus , we have rewritten ( [ rs1],[rs2 ] ) in terms of matrices obeying the usual algebra for the pasep ( [ pasep : de])([pasep : ew ] ) with redefined @xmath45 and @xmath11 . from this rewriting , the phase diagram can , in principle , be deduced using the general results of @xcite .
another discrete - time updating scheme is to update each site in a fixed sequence in each time step .
two particularly obvious choices of sequence are as follows .
[ [ forward - ordered - update ] ] forward - ordered update + + + + + + + + + + + + + + + + + + + + + + here , the timestep begins by updating site 1 wherein if site 1 is empty a particle enters with probability @xmath45 .
next , the bond @xmath637 is updated such that if site 1 is occupied and site 2 is empty the particle moves forward with probability @xmath26 , or else if site 2 is occupied and site 1 is empty the particle moves backward with probability @xmath96 .
then the bonds @xmath608 for @xmath638 are similarly updated in order .
the time step concludes with site @xmath14 being updated wherein if site @xmath14 is occupied the particle leaves the system with probability @xmath11 .
note that in the forward sequence it is possible for a particle to move several steps forward in one timestep .
[ [ backward - ordered - update ] ] backward - ordered update + + + + + + + + + + + + + + + + + + + + + + + in this case , the time step begins by updating site @xmath14 , then bonds @xmath639 to @xmath637 in backward sequence and finally site @xmath30 , where all the individual updates follow the usual rules , as described above .
note that due to particle - hole symmetry the dynamics of the vacancies in the backward order is the same as the updating of particles in the forward order .
therefore there is a symmetry between the steady states for the forward and backward orders . let us focus on the backward - ordered dynamics .
we construct the transfer matrix for the whole timestep as a product of operators for each update of the sequence @xmath640 where @xmath641 are as in ( [ trep],[lrrep ] ) .
the matrix product solution is of the usual form @xmath642 the cancellation mechanism is precisely the same as for the sublattice parallel case ( [ slcancel ] , [ slbccancel ] ) . in the ordered case
the hat matrices do not appear in the steady state weights at the end of the time step , rather they are auxiliary matrices which appear during the update procedure and move from right to left through the lattice .
we conclude that backward ordered updating has the exact same phase diagram as sublattice updating , although exact expressions for correlation functions do depend on the details of the updating@xcite .
finally the phase diagram for forward ordered updating can be obtained from the particle - hole symmetry mentioned above .
current and density profiles have been calculated in @xcite . in parallel dynamics ( sometimes referred to as fully parallel to distinguish from the sublattice updating ) _
all _ bonds and boundary sites are updated simultaneously .
this dynamics is considered the most natural for modelling traffic flow @xcite . in an update at the bond @xmath643 , if site @xmath16 is occupied and site @xmath36 is empty the particle moves forward with probability @xmath26 . under this type of updating scheme one can not include reverse hopping without introducing the possibility of conflicts occuring , or of a particle hopping to two places at once .
also note that it is the occupancies at the beginning of the timestep which determine the dynamical events . in this case
the matrix product solution is of the usual form ( [ mss ] ) but the algebraic relations have a more complicated structure which was first elucidated in @xcite .
there it was found that recursion relations between systems of different sizes were higher than first order : for example the relation @xmath644 which relates the weights of size @xmath14 to those of size @xmath486 and size @xmath645 , was discovered to hold .
this in turn implies that algebraic relations between the operators are _ quartic _ rather than quadratic .
for example the rules in the bulk are @xmath646 and there are other rules which we do not quote here for matrices near to the boundary ( see @xcite ) .
these rules were proved using a domain approach related to , but more complicated than , that presented in in section [ rrproof ] @xcite .
subsequently , an algebraic proof similar in spirit to the cancellation mechanism for the ordered sequential updating case was found @xcite but it is still too involved to present here . the quartic algebra , ( [ bulk1])([bulk4 ] ) , as well as the other conditions mentioned above , can be reduced to quadratic rules by making a convenient choice for the operators involved .
the trick is to write @xmath647 where @xmath648 , @xmath649 , @xmath650 , and @xmath651 are matrices of , in general , infinite dimension ; that is , @xmath567 and @xmath568 are written as rank four tensors with two indices of ( possibly ) infinite dimension and the other two indices of dimension two .
correspondingly , we write @xmath652 and @xmath653 in the form @xmath654 where @xmath655 , @xmath656 , @xmath657 , and @xmath658 are vectors of the same dimension as @xmath648 and @xmath650 .
@xmath648 , @xmath650 , @xmath655 , and @xmath659 satisfy the quadratic relations @xmath660 \ ; , \label{d1e1con}\]]@xmath661 and @xmath649 , @xmath651 , @xmath656 , and @xmath662 satisfy @xmath663 , \label{e2d2con}\]]@xmath664 and @xmath655 satisfying ( [ d1e1con],[d1v1 ] ) are presented in @xcite . in addition , one can choose @xmath665 , @xmath666 , @xmath667 , @xmath668 so that ( [ e2d2con ] ) , ( [ e2v2 ] ) reduce to ( [ d1e1con ] ) , ( [ d1v1 ] ) . along the curve @xmath669 scalar representations of @xmath648 , @xmath650 , @xmath659 and @xmath670 can be found and @xmath671 are @xmath672 operators . the continuous time limit is recovered by setting @xmath673 , replacing @xmath674 , @xmath675 and letting the time step @xmath676 . then ( [ d1e1con],[d1v1 ] ) reduce to the usual asep quadratic algebra ( [ int : de])([int : ew ] ) . the phase diagram is presented in figure [ parpd ] .
it appears similar to the familiar continuous time phase diagram except that the transition lines from low density to maximal current and high density to maximal current are at @xmath677 and @xmath678 respectively . in the continuous time
limit described above one recovers @xmath679 and @xmath680 .
also , in the limit of deterministic bulk dynamics , @xmath681 @xcite the maximal current phase disappears from the phase diagram as is also the case in other updating schemes in this limit .
[ parpd ] , @xmath682 .
mc is the maximum current phase , ld and hd are the low and high - density phases , respectively .
the straight dashed lines are the boundaries between the low density and maximal current phases and the high density and maximal current phase at @xmath677 and @xmath678 respectively .
the curved dashed line is the line given by ( [ specialline ] ) and intersects the line @xmath683 at @xmath684 ]
in this work we have reviewed the physical properties of systems that are driven out of equilibrium , with particular reference to those one - dimensional models that can be solved exactly using a matrix product method .
the most prominent physical phenomena exhibited by these systems are phase transitions which in open systems are induced by changes in the interactions with the boundaries , and in periodic systems by the addition of particle species that have different dynamics .
additionally we have seen in a number of cases the formation of shocks in these one - dimensional systems .
as we have also discussed , the matrix - product approach can be used to determine the steady - state statistics of models with a variety of microscopic dynamics .
the majority of models that have matrix product states involve diffusing particles with hard - core interactions and whose numbers are conserved , except possibly at boundary sites . under such conditions , certain generalisations to multi - species models have been found .
furthermore , there are a few cases in which non - conserving particle reactions can occur that also admit a convenient solution in terms of matrix products .
however , it has to be said that the dream scenario , of being able to systematically construct any nonequilibrium steady state in matrix product form starting from the microscopic dynamics of the system , appears a remote goal .
although the existence proofs discussed in section [ formal ] tell us that a matrix product formulation should nearly always be possible , we are in practice still feeling our way with particular examples .
there remain a number of simple physical systems that exhibit nontrivial out - of - equilibrium behaviour , but that have nevertheless been unyielding to exact solution by matrix products , or any other means .
it is perhaps appropriate here to highlight some of these cases as challenges for future research .
as was discussed in section [ pwdis ] , it is possible to solve for the steady state of the asep on the ring when particles have different hop rates , as would occur when each particle is randomly assigned a hop rate from some distribution . as far as we aware
, the corresponding dynamics with open boundaries has not been solved apart from the karimipour s model of overtaking which we reviewed in section [ sec : karimipour ] .
furthermore , if the disordered hop rates are associated with sites , rather than particles , even the model on the ring has so far eluded a complete exact solution .
indeed , even if only a single site has a different hop rate to the rest , the steady state does not appear to have a manageable form @xcite .
on the other hand , some nontrivial symmetries in the disordered case have been identified @xcite .
one phenomenon that can emerge when disorder is present is phase separation .
one sees this most clearly for the periodic system with particle - wise disorder , where as we noted in section [ bec ] , a condensation of particles behind the slowest may occur @xcite . with site - wise disorder @xcite , a flattening of the macroscopic current - density relation @xmath51
is seen ( recall that it is parabolic @xmath685 in the asep without disorder ) , and the densities at which transitions into a maximal current phase are correspondingly lowered @xcite ; it has been suggested that a linear portion in the @xmath51 curve may also provide a mechanism for phase separation @xcite .
another interesting observation is that under site - wise disorder the location of the first - order transition for the asep becomes sample - dependent ( i.e. , the free - energy - like quantity for the nonequilibrium system is not self - averaging ) @xcite . in all the models described in this work ,
particles occupy a single site of a lattice . if one has extended objects on a ring , the dynamics are unaffected . with open boundaries , however , the situation changes and even if there is a single particle species and all particles are the same length , a matrix product solution for the steady state has proved elusive .
although it is quite easy to arrange for cancellation of terms in the master equation corresponding to particles joining and leaving domains ( see section [ rrproof ] ) , the necessary cancellation with boundary terms is much harder to arrange : it is not even clear what the most convenient choice of boundary conditions should be ( e.g. , extended particles sliding on and off site - by - site , or appearing and disappearing in their entirety , or a mixture of the two ) .
a phenomenological diffusion equation for the macroscopic density profile in the open system with extended objects appearing and disappearing at both boundaries has been presented @xcite .
furthermore , it was conjectured in that work to be the equation that would result in the continuum limit ( along the lines of that taken in section [ sec : mfh ] ) , if an exact matrix - product solution of this model were to be found .
whether this should turn out to be the case or not , an application of the extremal current principle ( see [ sec : pd ] ) gives predictions for the phase diagram that are in good agreement with simulation data : the same three phases that emerge in the asep are seen , but with shifted transition points .
meanwhile an approach to this problem based on a local - equilibrium approximation @xcite has shown excellent agreement with monte carlo results and have also been conjectured to be exact .
the case of the open system with extended objects and a single spatial inhomogeneity in the hop rates has also been considered @xcite .
the bridge model @xcite comprises oppositely - charged particles in the sense of section [ 2species ] .
positive charges enter at the left at rate @xmath45 , hop with unit rate to the right and exit at the right boundary at rate @xmath11 .
the dynamics of negative charges is exactly the same , but in the opposite direction .
when two opposing particles meet they may exchange with some rate @xmath96 .
thus this model has both a parity and charge - conjugation symmetry . in certain parameter ranges ,
this symmetry is reflected in the steady state that results : both positive and negative charges flow freely in their preferred directions . in particular
, the limit @xmath686 at @xmath687 , for which an exact solution in terms of matrix products exists , is included within this range .
on the other hand , when @xmath11 is small the model s symmetries are found to be broken .
this has been proved in the @xmath688 limit @xcite but there is no known exact solution for general @xmath11 . in a finite system , the current flips between the positive and negative direction , but as the system size increases the time between flips increases exponentially , and so in the thermodynamic limit only the state with positive or negative current state is seen .
it is unclear whether the structure of the matrix product solution allows for the description of such symmetry - broken states .
this may also be related to whether the boundary conditions correspond to particle reservoirs with well - defined densities @xcite the abc model @xcite is another deceptively simple model that has yet to be completely solved .
this model has a ring that is fully occupied by particles from three species , which are labelled @xmath372 , @xmath373 and @xmath151 and exhibit a symmetry under cyclic permutation of the labels .
specifically , the rates at which neighbouring particles exchange are @xmath689 that is , if @xmath145 , @xmath372 particles prefer to be to the left of @xmath373s , @xmath373s to the left of @xmath151s and @xmath151s to the left of @xmath372s .
this model was discussed in section [ sec : abc ] and was shown to be exactly solvable when the numbers of @xmath372 , @xmath373 and @xmath151 particles are all equal as a consequence of detailed balance being satisfied . here
a phase separation is seen when @xmath690 . if @xmath170 , on the other hand , all configurations become equivalent and the stationary distribution is uniform .
when there is even the slightest imbalance and particle numbers _ and _ hop rates , a small current starts to flow and nothing is known about the properties of the stationary distribution .
the model is of particular importance because in the special case of equal numbers of each species , it is one of few models along with the ssep and kmp model @xcite where a free energy functional for the density profile is known .
we end by mentioning the original paradigm of a system driven out of equilibrium , the model due to katz , lebowitz and spohn @xcite .
this model has ising - like interactions between pairs of neighbouring spins and evolves by neighbouring spins exchanging places ( kawasaki dynamics ) but with a symmetry - breaking external field that favours hops of up - spins along a particular axis . when the system is open , or has periodic boundaries , a nonequilibrium steady state ensues . in the limit where spins become non - interacting and
there is only one spatial direction , the partially asymmetric exclusion process ( pasep ) described in section [ pasep ] is recovered .
when the spins interact , an ising - like steady state that involves products of @xmath672 matrices ( as described in section [ ising ] ) can be found , albeit only if certain conditions on the hop rates are satisfied .
( these conditions do admit nonequilibrium steady states in which a current flows , however ) . in two dimensions , one has , of course , a finite critical temperature in the absence of a driving force , and there has been particular interest in the nature of the low - temperature ordered phase in the presence of an external drive @xcite .
we discussed in section [ discretetime ] different updating schemes for which matrix product steady states have been determined . here
we mention some other updating schemes which have not been solved so far .
first an ordered discrete time sequential updating scheme could have the order as a randomly chosen sequence .
the forward and backward schemes of section refdiscretetime correspond to special sequences .
it would be of interest to construct the a matrix product steady state for an arbitrary fixed sequence .
recently a _ shuffled _ updating scheme has been considered @xcite , in which a new random sequence is chosen at each timestep .
it has been argued that this scheme is relevant to the modelling of pedestrian dynamics .
the shuffled updating scheme guarantees that each site or bond is updated exactly once in each timestep but is more stochastic than an ordered sequential scheme . again
it would be of interest to see if a matrix product could be used to describe the steady state .
we thank all the many colleagues with whom it has been a pleasure to discuss nonequilibrium matrix product states over the years . in particular
we acknowledge our collaborators and other authors whose work we have summarised . for their useful discussions and insightful comments during the preparation of this manuscript
we would wish to thank bernard derrida , rosemary harris , des johnston , joachim krug , kirone mallick , andreas schadschneider , gunther schtz and robin stinchcombe .
rab further acknowledges the royal society of edinburgh for the award of a research fellowship .
the continuity equation ( [ cty2 ] ) is a particularly simple case of a first order quasi - linear differential equation for the density @xmath691 the term on the right hand side would represent a source or sink of particles and is relevant for example to the case where creation and annihilation processes exist .
such an equation can generally be solved by the method of characteristics @xcite which involves identifying the characteristic curves along which information from the boundary or initial condition propagates through the space - time domain .
the characteristic curves satisfy @xmath692 of which the two independent conditions may be written @xmath693 these equations generally have two families of solution @xmath694 then the solution to ( [ cty3 ] ) is of the form @xmath695 where the function @xmath696 is fixed by the initial data . for the case ( [ cty2 ] ) , i.e. @xmath697 , the characteristics in the @xmath74@xmath6 plane are straight lines with slope @xmath59 and @xmath49 is constant along these lines , as can be seen from ( [ xrho ] ) .
more generally the characteristics are curves in the @xmath74@xmath6 plane and the density varies along the characteristic .
this can be seen by integrating ( [ xrho ] ) @xmath698 which gives @xmath49 implicitly as a function of @xmath6 , then the trajectory of the characteristic is given by @xmath699 if we consider a patch of the initial density profile of density @xmath700 at @xmath411 , its density evolves according to ( [ patden ] ) and ( [ patpos ] ) gives the position of the patch .
these more complicated characteristics can produce a richer for boundary induced phase transitions @xcite including stationary shocks .
in this appendix , we collect together general formul for obtaining both exact and asymptotic expressions for the coefficients appearing in a generating function if the latter is known .
we refer the reader to @xcite for proofs and more detailed discussion . for brevity , we introduce the notation @xmath701 to mean `` the coefficient of @xmath702 in the ( formal ) power series @xmath201 '' . [ [ lagrange - inversion - formula ] ] lagrange inversion formula + + + + + + + + + + + + + + + + + + + + + + + + + + if a generating function @xmath201 satisfies the functional relation @xmath703 where @xmath704 satisfies @xmath705 , one has an expression for @xmath74 as a function of @xmath57 which can be inverted to find @xmath57 as a function of @xmath74 .
this procedure then allows one to determine with relative ease the coefficient of @xmath702 in the expansion of any _ arbitrary _ function @xmath706 .
the formula for doing so reads @xmath707 \;.\ ] ] [ [ asymptotics ] ] asymptotics + + + + + + + + + + + if a generating function @xmath201 has a singularity at a point @xmath411 in the complex plane , it contributes a term @xmath708 to the coefficient @xmath701 , where @xmath372 and @xmath470 are constants to be determined . for sufficiently large @xmath22 , this coefficient is dominated by the singularity closest to the origin . to be more precise
, we say that two sequences @xmath709 and @xmath710 are asymptotically equivalent , denoted @xmath711 , if @xmath712 then , we have that @xmath713 .
let us decompose the generating function @xmath201 into its regular and singular parts around @xmath411 , denoting the former by @xmath714 .
that is , @xmath715 the cases that interest us are poles @xmath716 and algebraic singularities ( noninteger @xmath501 ) . for the former case one
finds @xmath717 for sufficiently large @xmath22 , the binomial coefficient behaves as @xmath718 replacing the factorial with a gamma function gives the extension of ( [ gf : pole ] ) to algebraic singularities @xmath719 the prefactor @xmath720 in this expression can be obtained by taking the limit @xmath721
in order to show that the integral representation of the normalisation ( [ asep : z0 ] ) is equivalent to the summation ( [ open : z2 ] ) , the simplest approach is to show that the generating functions for @xmath452 computed from the two expressions are equal . as we have seen in section [ gfa ] the generating function computed from the finite sum ( [ open : z2 ] ) yields ( [ open : zed ] ) which can be developed further to give @xmath722}{2 \left[z-\alpha(1-\alpha)\right ] \left[z-\beta(1-\beta)\right]}\;.\label{open : zed2}\end{aligned}\ ] ] we now turn to the integral representation ( [ asep : z0 ] ) which may be recast as a contour integral by change of variable @xmath723 @xmath724 where @xmath725 is the ( positively oriented ) unit circle in the complex @xmath422 plane .
we compute the generating function ( [ gendefa ] ) using ( [ zinta ] ) . summing a geometric series ( which converges for @xmath726 ) we find @xmath727 we now use the residue theorem to evaluate this integral .
we take @xmath105 and @xmath728 ( although we get the same final result for other ranges of @xmath105 and @xmath106 ) . in this case
the singularities within the unit circle are simple poles at @xmath729 , @xmath730 and @xmath731 where we define @xmath732 the residues from the three poles yield @xmath733 } -\frac{(1-b^2)}{2(a - b ) } \frac{\beta^2}{\left[z-\beta(1-\beta)\right]}\\ & & -\frac{(1-ab)(2-u_-^2-u_-^{-2 } ) } { 2z(1+a^2-a(u_-+u_-^{-1}))(1+b^2-b(u_-+u_-^{-1}))(u_- - u_+)}\;.\end{aligned}\ ] ] some simple algebra then shows that this expression is equivalent to ( [ open : zed2 ] ) .
here , we collect together some representations of the various matrix product algebras that have been discussed in this work .
we begin by recapping representations for the pasep algebra in the case of injection at the left boundary and extraction at the right boundary @xmath734 these representations are generalisations of the three representations for the totally asymmetric case first given in @xcite .
from section [ alsalam ] we have @xmath735 where @xmath736 and the boundary vectors are @xmath737 from this representation one sees that for certain parameter curves , namely @xmath738 that the representations become finite dimensional since @xmath739 and the upper left corner of the matrices become disconnected from the rest .
curves of this type were first noted in the more general case @xmath740 in @xcite and finite dimensional representations were catalogued in @xcite . in the limit
@xmath259 ( [ pasep : d1 ] ) and ( [ pasep : d1 ] ) have well - defined limits @xmath741 where @xmath742 .
a second representation of ( [ pasep : dea][pasep : ewa ] ) given in section [ pasymp ] is @xmath743 with @xmath744 where the parameters @xmath105 and @xmath106 are given in ( [ abcndef ] ) , and @xmath108 is a constant usually chosen so that @xmath109 . for these matrices , the @xmath259 limit is not defined .
a third representation , which we have not yet encountered , is @xmath745 where @xmath746 in the @xmath259 limit the elements of @xmath567 reduce to @xmath747 we now give representations corresponding to the physically relevant and nontrivial algebras classified in section [ 2species ] and the multispecies generalisations of section [ sec : multispecies ] .
[ [ solution - aii - and - multispecies - generalisation ] ] solution aii and multispecies generalisation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this is a multispecies model with matrices @xmath568 and @xmath538 , with an algebra @xmath748 where @xmath749 .
one possible representation is @xmath750 this is the generalisation of the third pasep representation ( [ de3 ] ) in the case @xmath137 .
[ [ solution - bi - and - bii ] ] solution bi and bii + + + + + + + + + + + + + + + + + + + with the algebra ( [ bialg ] ) , @xmath751 one has a representation @xmath752 where @xmath753 as noted in the main text a representation for solution bii is to use a presentation of the pasep algebra and write @xmath754 , i.e. a projector .
[ [ solution - cii - and - multispecies - generalisation ] ] solution cii and multispecies generalisation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this is the case of the asep with disordered hopping rates .
the matrix algebra ( [ msalg1 ] ) @xmath755 can be represented via @xmath756 where @xmath757 [ [ solution - d - and - deformed - commutators ] ] solution d and deformed commutators + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as noted in [ sec : closeg ] the solution class d operators can be constructed from tensor products of operators obeying deformed commutator relations .
the deformed commutator is defined as @xmath758 where @xmath361 is a deformation parameter : @xmath360 corresponds to the usual commutator @xmath759=0 $ ] .
deformed commutators appear in the analysis of the pasep , and in some of the two - species models of section [ 2species ] .
one possible representation of these matrices is @xmath760 we note that both @xmath355 and any product of @xmath355 and @xmath344 including at least one @xmath355 are traceless .
100 url # 1`#1`urlprefix zia r k p and schmittmann b 2007 probability currents as principal characteristics in the statistical mechanics of non - equilibrium steady states _ j. stat .
_ to appear ; cond - mat/0701763 leduc c , camps o , zeldovich k b , roux a , jolimaitre p , bourel - bonnet l , goud b , joanny j f , bassereau p and prost j 2004 cooperative extraction of membrane nanotubes by molecular motors _ proc .
natl . acad .
sci . _ * 101 * 17096 schtz g m 1996 experimental realizations of integrable reaction - diffusion processes in biological and chemical systems in _ proceedings of the @xmath765 nankai workshop on symmetry , statistical mechanics models , and applications _
, ed f y wu and m l ge ( world scientific , singapore ) godreche c , luck j m , evans m r , mukamel d , sandow s and speer e r 1995 spontaneous symmetry breaking : exact results for a biased random walk model of an exclusion process _ j. phys .
a : math . gen .
_ * 28 * 6039 | we consider the general problem of determining the steady state of stochastic nonequilibrium systems such as those that have been used to model ( among other things ) biological transport and traffic flow .
we begin with a broad overview of this class of driven diffusive systems which includes exclusion processes focusing on interesting physical properties , such as shocks and phase transitions .
we then turn our attention specifically to those models for which the exact distribution of microstates in the steady state can be expressed in a matrix product form .
in addition to a gentle introduction to this matrix product approach , how it works and how it relates to similar constructions that arise in other physical contexts , we present a unified , pedagogical account of the various means by which the statistical mechanical calculations of macroscopic physical quantities are actually performed .
we also review a number of more advanced topics , including nonequilibrium free energy functionals , the classification of exclusion processes involving multiple particle species , existence proofs of a matrix product state for a given model and more complicated variants of the matrix product state that allow various types of parallel dynamics to be handled .
we conclude with a brief discussion of open problems for future research . |
in 1969 , on the sixtieth anniversary of max born s @xcite first analysis of the field of a uniformly accelerated charge , ginzburg , nobelist in 2003 , reanalyzed @xcite this what he called``perpetual problem of classical physics , '' with the conclusion that the problem `` is already clear enough not to be regarded as perpetual . ''
ginzburg confirmed the presence of radiation and emphasized that the vanishing of the radiation reaction force during the uniformly accelerated motion of the charge `` is in no way paradoxical , in spite of the presence of radiation , '' since `` a non - zero total energy flux through a surface surrounding a charge at a zero radiation force is exactly equal to the decrease of the field energy in the volume enclosed by this surface . '' despite ginzburg s view , however , the problem does not seem to lose its a number of distinguished physicists who dealt with it before ginzburg like sommerfeld , schott , von laue , pauli and others have , after ginzburg , been followed by such authors as , for example , bondi @xcite , boulware @xcite , peierls @xcite , thirring @xcite and others @xcite .
the fields and radiation patterns from uniformly accelerated general multipole particles were also studied @xcite .
the december 2000 issue of annals of physics contains three papers by eriksen and grn @xcite with numerous references on .
( yet , except for @xcite and @xcite , the explicit citations above are not contained in @xcite . )
spacetimes describing play fundamental role in general relativity .
they are the only explicit solutions of einstein s field equations known which are radiative and represent the fields of finite sources . born fields in electrodynamics are produced by two charges moving along an in opposite directions with uniform accelerations of the same magnitude .
they have two symmetries : they are axially symmetric and symmetric with respect to the boosts along the axis of symmetry . their general - relativistic counterparts , the boost rotation symmetric spacetimes , are unique because of a theorem which roughly states that in axially symmetric , locally asymptotically flat spacetimes the only additional symmetry that does not exclude radiation is the boost symmetry .
the boost - rotation symmetric spacetimes have been used in gravitational radiation theory , quantum gravity , and as test beds in numerical relativity ; their general structure is described in @xcite , their applications and new references are given in the reviews @xcite .
one of the best known examples , the so - called c - metric , describing uniformly accelerated black holes , is the only boost - rotation symmetric solution known also for a nonvanishing cosmological constant @xmath0 .
asymptotically this c - metric approaches de sitter spacetime if @xmath1 .
it is well - known from the classical work of penrose @xcite on the asymptotic properties of fields and spacetimes that , in contrast to asymptotically minkowskian spacetimes with null ( lightlike ) conformal infinities @xmath2 , asymptotically de sitter vacuum spacetimes have two disjoint conformal infinities , past and future , which are both _
spacelike_. when @xmath3 , as in anti - de sitter space , the conformal infinity is timelike , and it is not disjoint .
( in the analytically extended c - metrics , there is an infinite number of such infinities which can be reached by going black holes like with a reissner - nordtrm black hole , but this is not pertinent to the present work . )
the importance of de sitter spacetime in the history of modern cosmology seems to grow steadily .
the de sitter universe became the standard cosmological model in steady state theory , more recently , as the of inflationary models , and today , with indications that @xmath1 in our universe , it is an asymptote of all indefinitely expanding friedmann - robertson - walker models with @xmath1 . in fact
much more general cosmological models with @xmath4 approach de sitter model asymptotically in time .
this manifestation of the validity of the @xcite , @xcite will also be noticed in the properties of the fields analyzed in this work .
motivated by the role of the born solution in classical electrodynamics , by the importance of the boost - rotation symmetric spacetimes in general relativity , and by the relevance of de sitter space in contemporary cosmology , we have recently generalized the born solution for scalar and electromagnetic fields to the case of two charges uniformly accelerated in de sitter universe @xcite . in the present paper
we give calculations and detailed proofs of the results and statements briefly sketched in our paper @xcite .
in addition , we investigate the character of the field in a number of various coordinate systems which are relevant either in a general - relativistic context or from a cosmological perspective .
the appropriate coordinates and corresponding tetrad fields were important in finding our recent results on a general asymptotic behavior of fields in the neighborhood of future infinity @xmath5 in asymptotically de sitter spacetimes @xcite . in obtaining these results
we were inspired by the inspection of the electromagnetic fields from uniformly accelerated charges in de sitter universe .
it was known from the work of penrose since late 1960 s that the radiation field is defined when @xmath5 is spacelike that it depends on the direction in which @xmath5 is approached .
however , no explicit models were available .
the investigation of the test fields of accelerated charges in de sitter universe has served as a useful example ; it was then generalized also to the study of asymptotic and radiative properties of the c - metric with @xmath1 @xcite , as well as to the case of the c - metric with @xmath3 when infinity is timelike @xcite .
( for other recent works on the c - metric , see , e.g. , @xcite . )
these studies led to more general conclusions @xcite : the directional pattern of gravitational and electromagnetic radiation near de sitter - like conformal infinity has a universal character , determined by the algebraic ( petrov ) type of a solution of the maxwell / einstein equations considered . in particular , the radiation field vanishes along directions opposite to principal null directions . very recently
analogous conclusions have been obtained for spacetimes with anti - de sitter asymptotics @xcite . since
past and future infinities are spacelike in de sitter spacetime , there exist particle and event horizons . under the presence of the horizons , purely retarded fields ( appropriately defined )
become singular or even can not be constructed at the , i.e. , at future light cones of the at @xmath6 at which the sources
the universe .
in @xcite we analyzed this phenomenon in detail and constructed smooth ( outside the sources ) fields involving both retarded and advanced effects .
as demonstrated in @xcite , to be is quite a different matter than to be .
this reveals the double meaning of the second perhaps somewhat enigmatic
part of the title of this paper .
its plan is as follows . in order to gain an understanding of the generalized born solution in de sitter space it is advantageous to be familiar with some details of the classical born solution in minkowski space .
hence , its properties most relevant for our purpose are summarized in section [ sc : bornm ] .
here we also discuss why in minkowski space problems with purely retarded fields of uniformly accelerated particles do not arise .
there exists vast literature on de sitter space in which various types of coordinates are employed .
we shall construct fields in de sitter space by using its conformal relations to minkowski space . for our aim coordinate systems on conformally compactified spaces and their properties
will be particularly useful .
these , together with several and coordinate systems , will be described and graphically illustrated in conformal diagrams in section [ sc : coor ] .
what is meant by is defined and the properties of the corresponding worldlines are studied in section [ sc : accpart ] . for technical reasons
it is more advantageous to consider particles which asymptotically start and end at the poles of coordinates covering de sitter space , i.e. , particles ( section [ ssc : patpoles ] ) . in order to find a direct relation between the standard form of the born solution produced by two charges at each time
located symmetrically with respect to the origin of minkowski space and the generalized born solution in de sitter space , it is necessary to construct also worldlines of uniformly accelerated particles which are ( section [ ssc : patequat ] ) . with the worldlines of accelerated particles available ,
it is advantageous to consider coordinates in de sitter space which are centered on these worldlines .
these and are obtained , in a constructive manner , in section [ sc : acccoor ] . section [ sc : fields ] is devoted to the fields from particles .
here we also study in detail their properties in various coordinate systems introduced before .
the fields of particles are found in section [ sc : bornds ] by a simple rotation .
starting from these fields we demonstrate by means of which limiting procedure the standard born field in minkowski space can be regained .
finally , we conclude by few remarks in section [ sc : concl ] . the paper contains a rather extensive appendix in which nine families of coordinate systems employed in the main text are described in detail , illustrated graphically , their relations are given , and corresponding metric forms as well as orthonormal tetrads are presented .
we believe the appendix can be used as a general - purpose catalogue in other studies of physics in de sitter spacetime .
it was einstein in 1908 , inspired by a letter from planck , who first defined a motion in special relativity @xcite .
a particle is in uniformly accelerated motion if its acceleration has a fixed constant value in instantaneous rest frames of the particle .
this can be stated in a covariant form ( see , e.g. , @xcite ) as @xmath7 @xmath8 being four - velocity , @xmath9 covariant derivative with respect to proper time , @xmath10 four - acceleration , and @xmath11 is the projection tensor into the hypersurface orthogonal to @xmath8 .
. implies @xmath12 so that the condition of uniform acceleration guarantees that the magnitude of the four - acceleration is constant , @xmath13 although @xmath14 . integrating eq . in minkowski spacetime
, one finds that the worldline of a uniformly accelerated particle is a hyperbola @xcite .
one can then choose an inertial frame , in which the initial three - velocity and three - acceleration are parallel ; in such frames the motion is spatially 1-dimensional .
it can be produced by putting a test charged particle into a homogenous electric field with initial velocity aligned with the field .
the motion along the @xmath15 axis is illustrated in fig .
[ fig : mink ] . there
, in fact , _ two _ particles uniformly accelerated in opposite directions are shown , the one moving along the positive ( @xmath16 for particle @xmath17 in the figure ) and the second one along the negative @xmath15 axis ( @xmath18 for particle @xmath19 ) ; their worldlines parametrized by proper time @xmath20 are @xmath21 or @xmath22 here we have chosen the particles to be at rest at @xmath23 at @xmath24
. then their three - acceleration at initial moment @xmath24 is @xmath25 . as @xmath26 , the three - velocity @xmath27 approaches the velocity of light
this is the well - known .
the worldlines of the particles coincide with the orbits of the boost killing vector in the @xmath28 plane , @xmath29 these orbits , given by @xmath30 , @xmath31 , are timelike at @xmath32 , but they are spacelike at @xmath33 . the fields ( scalar , electromagnetic , higher - spin ) produced by charged particles in the hyperbolic motion will have boost - rotational symmetry .
they are thus static in the region @xmath32 as introduced in @xcite , however , we can expect them to be radiative in the region @xmath33. consider a massless scalar field @xmath34 with the scalar charge source @xmath35 satisfying , in a general 4-dimensional spacetime , the wave equation @xmath36\,\sf=\sfsource\commae\ ] ] in which @xmath37 is the curved - space dalambertian , and @xmath38 is the scalar curvature ( of course , in minkowski space @xmath39 ) .
we are interested in a field due to two monopole particles with the same constant scalar charge of magnitude @xmath40 moving along hyperbolae . the source at a spacetime point @xmath41 is thus given by @xmath42 where @xmath43 denotes the worldlines of the particles .
the resulting fields may be written as @xmath44 where @xmath45 is produced by @xmath46 .
the retarded and advanced fields of these sources are constructed and analyzed in detail in ref .
it can be demonstrated that the retarded and advanced fields due to the particle @xmath17 or @xmath47 are all given by exactly _ identical _ expression @xmath48 which , however , is valid in different regions of spacetime .
namely , @xmath49 @xmath50 being the step function and upper / lower sign is valid for retarded / advanced case . the quantity @xmath51 in the denominator is given by @xmath52 it has the meaning of a or it is a spatial distance of the ( field ) point from the position of the source at retarded or advanced time . here , as usual , @xmath53 , @xmath54 , @xmath55
the fields , as well as , are , at first glance , axially ( rotationally ) symmetric .
they are also unchanged under the boost along the @xmath15 axis .
the field @xmath56 can , in fact , be viewed as the field due to both accelerated particles , i.e. , as the field corresponding to the source . inspecting regions at which the retarded and advanced fields are non - vanishing we discover that @xmath56 admits the interpretation as arising from 1-parametric combination of retarded and advanced effects from both particles : @xmath57 where @xmath58 is an arbitrary constant parameter . in particular ,
choosing @xmath59 , the field @xmath56 arises from @xmath60 from both particles . with @xmath61 ,
the field can be interpreted as being caused by purely retarded effects from particle @xmath17 in region @xmath62 , and by purely advanced effects from particle @xmath47 in region @xmath63 .
the case of electrodynamics is very similar .
the solution corresponding to the scalar field was found by born in 1909 @xcite .
it is customarily given in cylindrical coordinates ( see , e.g. , @xcite ) , however , in order to compare it with its generalization to de sitter universe , it is more convenient to write it down in spherical coordinates : @xmath64 the field can be obtained from the linard - wiechert retarded and advanced potentials of two charged particles moving along hyperbolae , however , in contrast to the scalar case when charges are exactly the same , the electric charges have _ opposite _ signs .
similarly to the scalar case , the field is smooth everywhere , except for the places where the particles occur .
@xmath65 can be interpreted in the precisely same way as the scalar field , i.e. , as the 1-parametric combination of retarded and advanced effects from both charges , analogously to eq . .
however , in the electromagnetic case an exact form of retarded and advanced fields from a single particle is a more subtle issue . considering that the field in the region @xmath62 may be interpreted as the retarded effect emitted from the charge which moves along @xmath66 , it is natural to try to exclude advanced effects of the other particle by requiring the field to vanish in the region @xmath63 ( cf .
[ fig : mink ] ) .
the field is then not smooth at the null hypersurface @xmath67 . in the scalar case
such a field _ does _ represent the pure retarded field of the single particle , cf .
. however , in the electromagnetic case the field @xmath68 corresponds to sources consisting not only of the particle but also of a moving along hypersurface @xmath69 with velocity of light @xcite .
nevertheless , it is possible to obtain a pure retarded field of the only single particle by modifying the field with a delta function valued term localized on @xmath69 @xcite . in de sitter space
such a modification is not feasible because the advanced fields can not be excluded .
the underlying cause is the _ null _
character of the past conformal infinity in minkowski spacetime , whereas in de sitter spacetime both future and past conformal infinities are _ spacelike_. as a consequence , the gauss constraint restricts the data at the spacelike past infinity , and it can be shown that a purely retarded field of a point - like charge can not satisfy this constraint @xcite .
the absence of purely retarded fields is also related to a different character of the past horizon of a particle . since the worldline of a particle the universe through the past spacelike infinity , there exists the past particle horizon , called also the . in de
sitter space a purely retarded electromagnetic field of a point - like charge can not be constructed on the whole cone .
in minkowski spacetime the creation light cone of a particle moving asymptotically in the past freely , coincides with the whole past null infinity , and thus it does not belong to the physical spacetime . eternally accelerated particles can the minkowski spacetime at a point of the past null infinity as , for example , uniformly accelerated particles do .
like in de sitter case , in conformal spacetime the past horizon of such particles forms the null cone but , in contrast to de sitter space , it has one generator in common with the null infinity . in physical spacetime
this horizon thus corresponds to a null hyperplane for the particle @xmath17 it is just the hyperplane @xmath69 ( cf .
[ fig : mink])and so its spatial sections are not compact .
thanks to this non - compactness the behavior of the retarded field on the horizon can be to the infinity .
we analyzed this issue in detail in ref .
the fields due to various types of uniformly accelerated sources in de sitter spacetime found in @xcite , as well as those described briefly in ref .
@xcite , were constructed by employing the conformal relation between minkowski and de sitter spacetimes .
when analyzing the worldlines of the sources in de sitter spacetime and their relation to the corresponding worldlines in minkowski spacetime we need to introduce appropriate coordinate systems .
suitable coordinates will later be used to exhibit various properties of the fields .
an extensive literature exists on various types of coordinates in de sitter space ( e.g. @xcite ) , but we want to survey some of them in this section . in particular , we relate them to the corresponding coordinates on conformally related minkowski spaces since this does not appear to be given elsewhere . in the next section , after identifying the worldlines of uniformly accelerated particles in de sitter space , we shall construct new coordinate systems tied to such particles , such as rindler - type coordinates , or robinson - trautman - type coordinates in which the null cones emanating from the particles have especially simple forms .
these coordinate systems will turn out to be very useful in analyzing the fields . here , in the main text , however , only a brief description of relevant coordinates will be given .
more details , including both formulas and illustrations , are relegated to the appendix .
as it is well - known from textbooks on general relativity ( for a recent pedagogical exposition , see @xcite ) , de sitter spacetime , which is the solution of einstein vacuum equations with a cosmological term @xmath70 , is best visualized as the 4-dimensional hyperboloid imbedded in flat 5-dimensional minkowski space .
it is the homogeneous space of constant curvature equal to @xmath71 .
hereafter , we use the quantity @xmath72 ( with the dimension of length ) to parametrize the radius of the curvature .
the spherical cosmological coordinates and a pair of uniformly accelerated particles @xmath73 and @xmath74 in de sitter universe : the conformal diagram ( above ) and projection on the spacelike cut @xmath75 in the standard cosmological spherical coordinates ( angle @xmath76 suppressed ) .
the whole de sitter spacetime could be represented by just the of the conformal diagram . for convenience ,
we admit negative values of radial coordinates and identify @xmath77 and @xmath78 ( see the text below eq . and the appendix ) .
, title="fig : " ] + the spherical cosmological coordinates and a pair of uniformly accelerated particles @xmath73 and @xmath74 in de sitter universe : the conformal diagram ( above ) and projection on the spacelike cut @xmath75 in the standard cosmological spherical coordinates ( angle @xmath76 suppressed ) .
the whole de sitter spacetime could be represented by just the of the conformal diagram . for convenience ,
we admit negative values of radial coordinates and identify @xmath77 and @xmath78 ( see the text below eq . and the appendix ) .
, title="fig : " ] the entire de sitter spacetime can be covered by a single coordinate system which we call @xmath79 , @xmath80 , @xmath81 , @xmath82 in which the metric reads @xmath83 clearly , we can imagine the spacetime as the time evolution of a 3-sphere which shrinks from infinite extension at @xmath84 to a radius @xmath85 , and then expands again in a time - symmetric way .
hence , we also call @xmath86 the .
the coordinate lines are shown in the conformal diagram , fig .
[ fig : ds ] .
the flat cosmological coordinates and particles @xmath73 , @xmath74 in de sitter space and in conformally related minkowski space .
the flat cosmological coordinates cover shaded region .
its boundary , @xmath87 , represents the horizon for observers at rest in these coordinates . ] in cosmology the most popular de sitter universe is obtained by considering only a half of de sitter hyperboloid foliated by flat 3-dimensional spacelike hypersurfaces labeled by timelike coordinate @xmath88 , cf .
[ fig : dsminkv ] . together with appropriate radial coordinate @xmath89 , the new coordinates , which we call , are given in terms of @xmath86 by @xmath90 implying the well - known metric @xmath91 these coordinates cover only of de sitter space as indicated by shading in fig .
[ fig : dsminkv ] .
de sitter introduced his model in what we call @xmath92 , @xmath93 ( see fig .
[ fig : dshyp ] ) related to @xmath86 by @xmath94 the metric @xmath95 shows that the time slices @xmath96 have the geometry of constant negative curvature , i.e. , as the standard time slices in an open frw universe
. the hyperbolic cosmological coordinates .
they cover only the shaded region and , therefore , only a part of the worldline @xmath73 .
the horizon @xmath97 arises for the observers who are at rest in the hyperbolic cosmological coordinates . ] the last commonly used coordinates in de sitter spacetime are @xmath98 , @xmath99 : @xmath100 covering also only a part of the universe .
the metric in these coordinates reads @xmath101 revealing that @xmath102 is a timelike killing vector in the region @xmath103 . among the coordinates introduced until now
only the standard coordinates @xmath104 cover the whole de sitter spacetime globally
. one can easily extend flat cosmological coordinates to cover ( though not smoothly ) the whole de sitter hyperboloid , which will be useful in discussion of the conformally related minkowski spacetime , cf .
we shall also use extensions of the static coordinates into the whole spacetime , using definitions , but allowing @xmath105 . in regions where @xmath106 coordinates @xmath107 and @xmath108 interchange their character , @xmath102 becomes a spacelike killing vector ( analogously to @xmath109 inside a schwarzschild black hole ) .
however , the static coordinates @xmath110 are not globally smooth and uniquely valued .
namely , @xmath111 at the cosmological horizons @xmath112 .
the static coordinates , extended to the whole de sitter space , are illustrated in fig .
[ fig : dsstat ] . here
we also indicate the regions in which @xmath113 is spacelike by bold @xmath114 ( ) and @xmath115 ( ) , whereas the regions in which it is timelike are denoted by @xmath116 ( containing the @xmath117 ) and @xmath118 ( containing the @xmath119 ) . hereafter , this notation will be used repeatedly .
the static coordinates and the wordlines of particles @xmath73 and @xmath74 .
these coordinates can be defined in the whole spacetime , however several coordinate patches , in diagram indicated by shaded and nonshaded regions , have to be used ( cf .
appendix [ asc : coorstatt ] and [ asc : coorstats ] ) .
these regions are separated by the cosmological horizons at @xmath120 , where @xmath121 .
the vector @xmath102 is a killing vector of de sitter spacetime .
it is timelike in the domains @xmath116 and @xmath118 ( shaded regions ) and spacelike in the domains @xmath114 and @xmath115 .
the histories of both particles @xmath73 and @xmath74 belong to the domains @xmath116 and @xmath118 . ] the conformal structure of minkowski and de sitter spacetimes , their conformal relation , and their conformal relation to various regions of the einstein static universe have been discussed extensively in literature ( see , e.g. , @xcite ) .
the complete compactified picture of these spacetimes , in particular the 3-dimensional diagram of the compactified minkowski and de sitter spaces @xmath122 as parts of the einstein universe represented by a _
solid _ cylinder can be found in @xcite .
we refer the reader especially to section iii of @xcite where we explain and illustrate the compactification in detail . in the present paper
we shall confine ourselves to the 2-dimensional penrose diagrams .
the basic covering globally de sitter spacetime including the conformal infinity are simply related to the standard coordinates as follows : @xmath123 @xmath124 , @xmath125 .
the metric becomes @xmath126 demonstrating explicitly the conformal relations of de sitter spacetime to the einstein universe : @xmath127 therefore , we also call coordinates @xmath128 the
. the conformal diagram of de sitter spacetime is illustrated in fig .
[ fig : ds ] .
the past and future infinities , @xmath129 and @xmath130 are spacelike , the worldlines of the north and south poles ( given by the choice of the origin of the coordinates ) are described by @xmath131 and @xmath78 .
the whole de sitter spacetime could be represented by just the of fig .
[ fig : ds ] .
indeed , it is customary to draw this half only and to consider any point in the figure as a 2-sphere , except for the poles @xmath132 .
as we shall see , the formulas relating coordinates on the conformally related de sitter and minkowski spacetimes have simpler forms if we admit negative values of the radial coordinate @xmath133 covering the left half of the diagram .
we shall thus consider the diagrams as in fig .
[ fig : ds ] to represent the cuts of de sitter spacetime along the axis going through the origins ( through north and south poles
analogously to the cuts along the @xmath15 axis in @xmath134 ) . the axis , i.e. , the main circle of the spatial spherical section of de sitter spacetime , is typically chosen as @xmath135 .
thus , in the diagram the point with @xmath136 , @xmath137 , @xmath138 is identical to that with @xmath139 , @xmath140 , and @xmath141 .
we use the same convention also for other radial coordinates appearing later , as explicitly stated in the appendix ( cf .
also appendix in @xcite ) .
we admit negative radial coordinates only when describing various relations between the coordinate systems . in the expressions for the fields in the following sections
only positive radial coordinates are considered .
as mentioned above , in @xcite we constructed fields on de sitter spacetime by conformally transforming the fields from minkowski spacetime .
now can be used in the conformal relation to de sitter space , depending on which region of a minkowski space is mapped onto which region of de sitter space .
consider , for example , minkowski space with metric @xmath142 given in spherical coordinates @xmath143 .
identify it with de sitter space by relations @xmath144 the inverse relation is given in the appendix . in the coordinates
@xmath143 the de sitter metric becomes @xmath145 so that @xmath146 the coordinates @xmath143 can , of course , be used in both de sitter and minkowski spaces .
[ fig : dsminkv ] illustrates the coordinate lines .
it also shows how four regions i , ii , iii , and iv of minkowski space are mapped onto four regions of de sitter space by relations .
we call @xmath147 since their radial coordinate @xmath148 coincides with that of the flat cosmological coordinates and the time coordinate is simply related to @xmath149 as @xmath150 the caron or the check ( still better ) formed by cosmological horizon at @xmath151 in de sitter space ( cf.fig . [ fig : dsminkv ] ) inspired our notation of these coordinates .
it is possible to introduce analogously the coordinates @xmath152 given in the appendix , eqs .
, , that cover nicely the past conformal infinity but are not smooth at the cosmological horizon @xmath153 ; in this case they form the hat in the conformal diagram ( see fig . [
fig : coorcma ] in the appendix ) . from relations
it is explicitly seen why , when writing down mappings between de sitter and minkowski spaces and drawing the corresponding 2-dimensional conformal diagrams , it is advantageous to admit negative radial coordinates .
if we would restrict all radial coordinates to be non - negative , we would have to consider the second relation in eq . with different signs for regions iii and ii in de sitter space : in iii @xmath154 , but in iii we would have @xmath155 .
the conformally minkowski coordinates .
they cover the whole conformally related minkowski space but only a part of corresponding de sitter space .
this minkowski space is related to that in fig .
[ fig : dsminkv ] by a shift by @xmath156 in the direction of the conformally einstein coordinate @xmath157 . ]
another mapping of minkowski on de sitter space will be used to advantage in the explicit manifestation that the generalized born solution in de sitter space goes over to the classical solution . instead of the mapping ,
consider the relations @xmath158 ( see eq .
for the inverse mapping ) , which turn the metric into @xmath159 we again obtain the de sitter metric in the form explicitly conformal to the minkowski metric with , however , a different conformal factor from that in eq . :
@xmath160 ( for the use of the de sitter metric in form in the work on the domain wall spacetimes , see @xcite . )
the relation of minkowski space to de sitter space based on the mapping is illustrated in fig .
[ fig : dsminko ] .
clearly , the minkowski space in this figure is shifted by @xmath156 in @xmath161 coordinate , as compared with minkowski space in fig .
[ fig : dsminkv ] ( eq . ) . indeed , replacing @xmath161 by @xmath162 in eq .
, we get @xmath163 , @xmath164 with @xmath165 given by eq . .
since coordinates @xmath166 are not connected directly with any cosmological model and correspond to minkowski space on de sitter space ( fig .
[ fig : dsminko ] ) , we just call them . in ref .
@xcite still another minkowski space is related to de sitter space
one which is shifted in @xmath161 coordinate by another @xmath156 . as mentioned below eq .
, the cosmological horizon forms hat in this case and the corresponding coordinates are accordingly denoted as @xmath152 .
they are given explicitly in appendix [ asc : coorcma ] and fig .
[ fig : coorcma ] . the three sets of coordinates @xmath147 , @xmath167 , and @xmath152 ( with the same @xmath168 ) relating naturally minkowski spaces to de sitter space are suitable for different purposes .
the third set describes conveniently the past infinity of de sitter space that is why it was used extensively in @xcite where we were interested in how the sources enter ( are ) de sitter universe .
the second set will be needed in section [ sc : bornds ] for exhibiting the flat - space limit of the generalized born solution .
the first set describes nicely the future infinity and will be employed when analyzing radiative properties of the fields . with all the coordinates discussed above , corresponding double null coordinates
can be associated ; some of them will also be used in the following .
their more detailed description and illustration is presented in section [ asc : coornull ] of the appendix . before concluding this section let us notice that the observers which are at rest in cosmological coordinate systems @xmath86 , @xmath169 , and @xmath170 move along the geodesics with proper time @xmath171 , @xmath149 , and @xmath172 respectively .
these geodesics are also the orbits of the conformal killing vectors .
indeed , the symmetries of minkowski spacetime and of the einstein universe become conformal symmetries in conformally related de sitter spacetime . in particular
, we shall employ the fact that since @xmath173 and @xmath174 are timelike killing vectors in minkowski spacetime and @xmath175 is a timelike killing vector in the einstein universe , the vectors @xmath176 are timelike conformal killing vectors in de sitter spacetime . as mentioned below eq .
, @xmath113 is a killing vector which is timelike for @xmath177 .
in section [ sc : bornm ] we defined uniformly accelerated motion in minkowski spacetime . however , the formulas given there , being in covariant forms , remain valid in de sitter spacetime . as explained in @xcite in detail , a simple way of obtaining a worldline of a uniformly accelerated particle in de sitter spacetime is to consider a suitable particle moving with a _ uniform velocity _ in minkowski spacetime and use the conformal relation between the spaces .
the worldlines of uniformly accelerated charges .
the particles @xmath73 and @xmath178 start and end at the @xmath74 , @xmath179 start and end at the south pole .
particles @xmath178 , @xmath179 have a higher magnitude of acceleration @xmath180 than particles @xmath73 , @xmath74 .
they are characterized by a negative parameter @xmath181 , whereas particles @xmath73 , @xmath74 have a positive @xmath182 . ]
consider a particle moving with a constant velocity of magnitude @xmath183 such that for @xmath184 it moves in a negative direction along the @xmath185 axis of the inertial frame in minkowski space @xmath186 with coordinates @xmath143 and passes through @xmath187 at @xmath188 : @xmath189 substituting into transformation , we find @xmath190 or expressing minkowski proper time @xmath191 in terms of the proper time of de sitter spacetime , @xmath192 we obtain @xmath193 here @xmath194 , @xmath195 takes values such that @xmath124 and @xmath125 for @xmath196 , or @xmath133 for @xmath197 .
upper sign is valid for the particle starting and ending with @xmath198 ( particle @xmath73 in fig .
[ fig : wldln_p ] ) , lower sign for the particle starting and ending at @xmath199 ( particle @xmath200 in fig . [ fig : wldln_p ] ) .
one can make sure by direct calculations of the four - acceleration ( for its simplest form in the static coordinates , see below ) that these worldlines describe the uniformly accelerated motion as defined in section [ sc : bornm ] , the magnitude of the acceleration being @xmath201 since de sitter universe represents the asymptotic state of all three types of indefinitely expanding frw models with @xmath202 , it is of interest to find out the form of these worldlines in the three types of cosmological frames spherical , flat , and hyperbolic introduced in section [ sc : bornm ]
. in terms of cosmological spherical coordinates the worldlines are given by @xmath203 in flat cosmological coordinates , which cover only half of de sitter space , we obtain just particle @xmath204 described by the worldline @xmath205 finally , in hyperbolic cosmological coordinates , which are also not global , we obtain again one particle s worldline only given in terms of its proper time as @xmath206 these formulas have no meaning for @xmath207 where the inverse hyperbolic functions are not defined .
this corresponds to the fact that for such @xmath208 the particle occurs in the region where the hyperbolic cosmological coordinates are not defined ( cf .
[ fig : dshyp ] ) . excluding the proper time we find the worldlines to be given by remarkably simple formulas in the three systems of the cosmological coordinates :
it is of interest to see what are the physical radial velocities which will be observed by three types of the fundamental cosmological observers , i.e. , those with fixed @xmath210 , @xmath148 , and @xmath211 , respectively , whose proper times are @xmath171 , @xmath149 , and @xmath212 , respectively .
such velocities can be defined by the covariant expression @xmath213{1}\ , \frac{d\prt_\ds}{d\prt_\obs}\commae\ ] ] where @xmath214 is the particle s four - velocity , @xmath208 its proper time , @xmath215{1}}$ ] is the unit spacelike vector in the direction of the radial coordinate @xmath216 , and @xmath217 , respectively , i.e. , in directions @xmath218 , @xmath219 and @xmath220 , and @xmath221 is the proper time of an observer , i.e. , @xmath222 or @xmath172 , respectively .
since all three cosmological metrics are diagonal the expression takes on the form @xmath223 the results are of interest : @xmath224 consider first the picture in spherical cosmological coordinates , eqs . and .
only in this frame both particles are present .
they start asymptotically at antipodes of the spatial section of de sitter space at @xmath6 ( @xmath225 ) and move one towards the other until @xmath226 , the moment of maximal contraction of de sitter space ( of de sitter hyperboloid ) , when they stop , @xmath227 .
then they move , in a time - symmetric manner , apart from each other until they reach future infinity asymptotically at the antipodes from which they started .
in contrast to the flat space case , the particles do not approach the velocity of light in this global spherical cosmological coordinate system , the asymptotical magnitude of their velocity being equal to @xmath228 ( cf .
hence , curiously enough , the particles approach the antipodes asymptotically with a finite nonvanishing velocity ( for an intuitive insight into this effect , see below ) .
although the particles @xmath204 and @xmath200 do not approach infinities with velocity of light , they _ are _ causally disconnected as the analogous pair of particles in minkowski space ( cf .
[ fig : mink ] and fig .
[ fig : wldln_p ] ) .
no retarded or advanced effects from the particle @xmath204 can reach the particle @xmath200 and vice versa .
next , consider flat and hyperbolic observers .
as seen from eq . , with respect to the flat cosmological coordinates the particle @xmath73 moves with the same velocity @xmath228 all the time .
and the same velocity is asymptotically , at @xmath229 , reached by this particle in the hyperbolic cosmological coordinates .
the magnitude of the asymptotic values of the velocity at @xmath5 is , in fact , equal to the velocity of the particle in minkowski space from which we constructed uniformly accelerated worldlines by a conformal transformation .
the identity of all these velocities is understandable : the magnitude of the velocity with respect to an observer can be determined by projecting the particle s four - velocity on the observer s four - velocity , i.e. , by the angle between these directions . in de
sitter space all three types of cosmological observers reach @xmath5 with the same four - velocity ; moreover , this four - velocity is at @xmath5 identical to the four - velocity of observers at rest in conformally related minkowski space .
but a conformal transformation preserves the angles and thus , the velocities with respect to the three types of cosmological observers in de sitter space and the velocity in the conformally related minkowski space must all be equal given by the angle @xmath181 .
it is worth noticing yet what is the _ initial _ velocity of the particle @xmath73 in hyperbolic cosmological coordinates . regarding fig .
[ fig : dshyp ] we have @xmath230 , @xmath231 at the of the particle at @xmath6 . from eq .
we get @xmath232 which in the magnitude is the same as in spherical cosmological coordinates but has opposite sign since the particle moves in the direction of increasing negative @xmath217 .
more interesting is how the particle enters the upper region of the hyperbolic coordinates .
[ fig : dshyp ] suggests that its velocity must approach the velocity of light since at this boundary the fundamental observers of the hyperbolic cosmological frame themselves approach the velocity of light . indeed
, at this boundary @xmath233 , @xmath234 , and the expression implies @xmath235 . by far the simplest description of the particles
is obtained in the static coordinates @xmath110 . using , for example , the relation @xmath236 ( cf .
eqs . , ) , and substituting from eq . , we find that the worldlines of both particles @xmath204 and @xmath200 are given by remarkably lucid forms @xmath237 these expressions imply that the four - acceleration @xmath238 is simply @xmath239 where @xmath240 is a unit spatial vector in the direction @xmath241 of the static radial coordinate @xmath242 , and we introduced constant @xmath243 which represents the value of the acceleration of the particles .
we thus find the uniformly accelerated particles in de sitter spacetime to be at rest in the static coordinates at fixed values @xmath244 of the radial coordinate .
two charges moving along the orbits of the boost killing vector in minkowski space are at rest in the rindler coordinate system and have a constant distance from the spacetime origin , as measured along the slices orthogonal to the killing vector .
similarly , we see that the worldlines @xmath73 and @xmath74 are the orbits of the static killing vector @xmath113 of de sitter space .
the particle @xmath73 ( respectively , @xmath74 ) has , as measured at fixed @xmath245 , a constant proper distance from the origin @xmath246 ( @xmath226 ) , @xmath131 ( respectively , @xmath78 ) . as with rindler
coordinates in minkowski space , the static coordinates cover only a of de sitter space . in the other
half the killing vector becomes spacelike . owing to caused by the presence of @xmath247 , fundamental cosmological observers moving along geodesics @xmath248 constant are one from the others .
their initial implosion starting at @xmath225 is stopped at @xmath226 and changes into expansion .
clearly , a particle with constant @xmath244hence a constant proper distance from the particle at @xmath249must be accelerated towards that particle . in eq .
we have denoted the radial tetrad component of the acceleration in the static coordinates by @xmath250 ; notice that , in contrast to the magnitude of the acceleration @xmath251 ( cf .
eq . ) , @xmath250 can be negative as , in fact , it is the case with _ both _ particles @xmath73 and @xmath74 , assuming that the static radial coordinate of the particles is positive , @xmath252 .
geometrically , the four - vectors of the acceleration of the particles point in opposite directions towards @xmath117 , the other towards @xmath119 . since , however , one needs two sets of the static coordinates to cover both particles , and the radial coordinate @xmath242 increases from both @xmath117 and @xmath119 worldlines ( cf . fig .
[ fig : dsstat ] ) , the accelerations of both particles point in the direction of decreasing @xmath242 s and is thus negative .
all the particles we are considering perform 1-dimensional motion only , hence we use for the description of their worldlines the same convention as for the 2-dimensional diagrams with time and radial coordinates we allow the radial coordinate to take negative values .
thus , for example , consider a particle with worldline @xmath253 which is a of the worldline @xmath73 with respect to @xmath131 ( see fig .
[ fig : wldln_p ] ) .
the particle @xmath253 moves in the region of negative @xmath254 , respectively @xmath242 , it has an acceleration positive , @xmath255 ( i.e. , @xmath197 ) , and its four - acceleration vector is pointing in the direction of increasing @xmath242 . with our convention ,
the particle @xmath253 is just that which moves from @xmath117 along the @xmath256 direction .
this convention will be particularly useful when we shall construct worldlines of uniformly accelerated particles which start and end at the equator .
those which move in the region @xmath257 will have negative @xmath250 , those moving with @xmath258 will have positive @xmath250see section [ ssc : patequat ] . an intuitive geometrical understanding of the worldlines of uniformly accelerated particles in de sitter spacetime can be gained by considering de sitter space as a 4-dimensional hyperboloid @xmath259 in 5-dimensional minkowski space .
the spherical cosmological coordinates @xmath104 are then identical to the hyperspherical coordinates on this hyperboloid .
the worldlines of the north and south poles , @xmath260 , can be obtained by cutting the hyperboloid by a timelike 2-plane @xmath261 , given by @xmath262 .
the worldlines of our uniformly accelerated particles @xmath73 and @xmath74 then arise when the hyperboloid is cut by a timelike 2-plane @xmath263 parallel to @xmath261 at a distance @xmath264 from the origin @xcite .
@xmath263 is thus given by @xmath265 , @xmath266 . from the definition of the hyperspherical
coordinates it follows @xmath135 and @xmath267 , i.e. , @xmath268 , which is just eq . describing @xmath73 and @xmath74 . from this construction , the curious result mentioned above that @xmath73 and @xmath74 approach antipodes @xmath117 and @xmath119 asymptotically with a fixed speed @xmath228 in spherical cosmological coordinates is not so surprising :
thanks to the expansion of de sitter spacetime all fundamental cosmological observers with arbitrarily small @xmath269 will , in the limit @xmath270 , eventually cross the plane @xmath263 , and thus the particle @xmath73 ; however at any finite but arbitrarily large @xmath271 there will be observers with @xmath272 which are still moving towards the particle @xmath73 . the same , of course , is true with the symmetrically located particle @xmath74 and corresponding observers close to @xmath119 .
the worldlines of uniformly accelerated charges located symmetrically with respect to the origin ( north pole ) of the standard spherical coordinates in de sitter space .
the particles and at the equator .
they are causally disconnected as a corresponding pair in minkowski space ( cf .
[ fig : mink ] ) .
the value @xmath273 of the acceleration of these particles is positive ( cf . the version of eq . ) . ]
the rotated spherical coordinates @xmath274 on 3-sphere ( the cut @xmath275 ) .
the relation between the coordinates is given in eq . .
] in the classical born solutions both charges are , at all times , located symmetrically with respect to the origin of the minkowski coordinates ( see fig .
[ fig : mink ] ) . in order to demonstrate explicitly that a limiting procedure exists in which our generalized born s solution goes over to its classical counterpart
, we shall now construct the pair of uniformly accelerated particles which are , at all times , symmetrically located with respect to the origin of the standard spherical coordinates in de sitter space , i.e. , with respect to the at @xmath117 .
asymptotically at @xmath225 these two particles both start ( ) with the same speed at the equator , @xmath276 , at the antipodal points @xmath277 and @xmath256 . as the universe contracts , they both move symmetrically along the axis @xmath135 , reach some limiting value @xmath278 at the moment of time symmetry , and accelerate back towards the equator , reaching the initial positions asymptotically at @xmath279 .
these two particles are illustrated in fig .
[ fig : wldln_e ] , with their worldlines denoted by @xmath17 and @xmath47 . in fig .
[ fig : rotation ] , a snapshot at @xmath75 is depicted . comparing fig .
[ fig : wldln_e ] with fig .
[ fig : wldln_p ] , it is evident that the particles @xmath17 and @xmath47 are located with respect to the point @xmath276 , @xmath277 in exactly the same manner as the particles @xmath204 and @xmath200 are located with respect to the pole @xmath117 ( or , rather , as the particles @xmath280 , @xmath281 , since we chose @xmath17 , @xmath47 to have positive @xmath273 in fig .
[ fig : wldln_e ] ) . owing to the global homogeneity of de sitter space and the spherical geometry of its slices @xmath75 , the worldlines of the particles @xmath17 and @xmath47 can be constructed by a suitable rotation of the worldlines of the particles @xmath204 and @xmath200 . in section [ sc : bornds ]
the same rotation will be applied to obtain the fields of these particles we rotate the coordinates @xmath248 into new coordinates @xmath282 which , as a pole , have the point @xmath276 , @xmath277 ( see fig . [
fig : rotation ] ) .
the relations between these coordinates follow from the spherical geometry : @xmath283 the new worldlines , @xmath17 and @xmath47 , will then be given by eqs .
in which @xmath248 are replaced by rotated coordinates @xmath282 . substituting for these by using relations , we find the worldlines @xmath17 , @xmath47 in the original coordinates to be described by the expressions : @xmath284 with the values of @xmath285 from @xmath286 and upper ( lower ) sign corresponding to the particle starting at the positive ( negative ) value of @xmath210 , i.e. , to the particle @xmath17 ( or @xmath47 , respectively ) . excluding the proper time @xmath287 , we arrive at simple result ( cf .
@xmath288 as @xmath289 , then indeed @xmath290 ; at @xmath226 , @xmath291 , in agreement with the of the particles @xmath204 , @xmath200 from @xmath117 at @xmath226 .
in the spherical rescaled coordinates , eqs .
read @xmath292 again with the values of @xmath285 and @xmath293 from @xmath286 .
. becomes @xmath294 although the flat ( rescaled ) cosmological coordinates cover only parts of the worldlines @xmath17 , @xmath47 ( see figs .
[ fig : wldln_e ] and [ fig : dsminkv ] ) , we transcribe the equations above also into these frames in which the particles at @xmath295 at the cosmological horizon at @xmath87 .
we find @xmath296 so that eq
. translates into the relations @xmath297 as @xmath298 , we have @xmath299 , as it corresponds to @xmath300 ; at @xmath301 , we get @xmath302here the particles enter flat cosmological frame at the horizon ( cf . fig . [
fig : wldln_e ] ) .
the worldlines @xmath17 , @xmath47 are situated outside the regions covered by our choice of the hyperbolic cosmological coordinates .
similarly , we get only finite parts of @xmath17 , @xmath47 in our static coordinates .
of course , we could rotate the static coordinates to cover both particles but then we arrive at exactly the same picture as with the particles @xmath204 , @xmath200 considered above . our primary reason to discuss the pair @xmath17 , @xmath47 is to demonstrate explicitly how our fields go over into the classical born solution in the limit of vanishing @xmath0 . for this purpose , it will be important to have available also the description of the worldlines @xmath17 , @xmath47 in the minkowski coordinates introduced in eqs . .
as it is obvious from fig .
[ fig : dsminko ] , these coordinates cover both worldlines @xmath17 and @xmath47 completely . using the relations inverse to eqs . given in the appendix , eq . , we find eqs . to imply @xmath303 where @xmath304 and @xmath305 is the proper time measured in minkowski space @xmath306 related to de sitter space by conformal mapping , : @xmath307 consequently , @xmath308 which is the simplest form of the hyperbolic motion with the uniform acceleration @xmath309 as measured in minkowski space ( cf.eqs .
for the investigation of the radiative properties and other physical aspects of the fields , the use of ( physically equivalent ) particles @xmath204 , @xmath200 , i.e. , those of spherical coordinates is technically more advantageous .
we shall now return back and construct frames with the origins located directly on these particles .
in such frames , various properties of the fields will become more transparent than in the coordinates introduced so far .
as we have seen in the preceding section , the uniformly accelerated particles @xmath73 and @xmath74 are at rest in static coordinates @xmath110 at given @xmath310 , where @xmath311 is the magnitude of the acceleration . in order to investigate the properties of the fields , in particular , in order to see what is the structure of the field along the null cones with vertices at the particle s position , i.e. , what is the field by the particle at a given time , it is useful to construct coordinate frames centered on the accelerated particles .
such systems of coordinates are used to describe accelerating black holes in general relativity ( like c - metrics , known also for @xmath312 , cf.@xcite ) , so that their properties on de sitter background may indicate what is their meaning in more general cases in situations when they are centered on gravitating objects rather than on test particles .
we shall now describe three coordinate systems of this type : the , the , and the , all centered on the worldlines @xmath73 and @xmath74 . instead of writing down just the transformation formulas , we wish to indicate some steps how these coordinates can be obtained naturally .
we list only the main transformation relations here , many other formulas and forms of the metrics can be found in the appendix .
let us also note that in this section we assume @xmath313 , i.e. , @xmath314 , and we use only static radial coordinate with positive values , i.e. , @xmath315 .
we begin with the construction of .
this type of coordinates was recently introduced @xcite by another method in the context of the with @xmath70 . in the preceding section we obtained the worldlines @xmath73 , @xmath74 of uniformly accelerated particles in de sitter space by starting from a particle moving with a uniform velocity @xmath316 in a negative direction of the @xmath185 axis in the inertial frame @xmath143 in minkowski space @xmath186 which passes through @xmath187 at @xmath188 ( see eqs . , ) ; and we used then the conformal relation between minkowski and de sitter spaces to find @xmath73 , @xmath74 .
therefore , let us first construct a frame centered on the uniformly moving particle in @xmath186 . using spherical coordinates again , this boosted frame denoted by primes is related to the original one simply by @xmath317 the 2-dimensional conformal diagrams of de sitter space based on the static , non - accelerated coordinates ( upper diagram ) , and on the accelerated coordinates ( lower diagram ) .
starting from static coordinates @xmath318 , one can draw the conformal diagram of the axis @xmath135 in which the conformal past and future infinities , @xmath2 ( @xmath319 ) , are horizontal ( double ) lines .
in addition to static coordinates @xmath110 , also accelerated coordinates @xmath320 are indicated in both diagrams .
these have a coordinate singularity for @xmath321 ( drawn as a dashed line ) .
the origins of the accelerated coordinates , @xmath322 ( thick lines ) , are worldlines of uniformly accelerated particles . in the conformal diagram of the axis @xmath323 based on accelerated coordinates , the origins @xmath322 and the coordinate singularity @xmath321 of the accelerated frame are straight lines ; the true infinities @xmath324 have a upwards or , depending on the angle @xmath325 .
the hypersurface @xmath321 corresponds to the boosted hyperplane @xmath326 , whereas the conformal infinity corresponds to @xmath327 ( the relation of both hyperplanes can be well understood in the diagram of the conformally related minkowski space @xmath328 ) .
the diagrams in which the conformal infinities @xmath324 are _ not _ straight naturally arise in the studies of the c - metric with @xmath70 ( de sitter space being a special case of this class of the metrics)see @xcite .
in general , outside the axis @xmath329 , the transformations between the static and accelerated coordinates mix radial and angular coordinates @xmath330 and @xmath331 , as is seen also in the following fig . [
fig : accrth ] .
the sections @xmath332 ( for some general @xmath333 ) are also shown in fig .
[ fig : cooracc ] in the appendix . ]
the accelerated coordinates @xmath334 on the sections @xmath335 of de sitter space ( coordinate @xmath336 suppressed ) . in the region where @xmath337 is timelike ( @xmath338 ) , the cut @xmath335 is a spacelike sphere ( diagram a ) . in the region where @xmath337 is spacelike ( @xmath339 and @xmath340 )
, it is a timelike hyperboloid ( diagram b ) .
the diagrams are not in the same scale the radius of the sphere and of the neck of the hyperboloid should be the same .
the axis @xmath323 corresponds to the lines @xmath335 of fig .
[ fig : acc ] . the coordinate singularity @xmath341 is also indicated . for more details
see the text . ] the @xmath336-coordinate does not change and will be suppressed in the following . from here
@xmath342 the original frame @xmath343 in minkowski space @xmath186 is related to the static coordinates @xmath344 in de sitter space by ( cf . eqs .
, ) @xmath345 the metrics of the two spaces are related by @xmath346 , @xmath347 being given by eq .
cf.eq . .
now , let us introduce coordinates @xmath348 given in terms of @xmath349 by exactly the same formulas as coordinates @xmath350 are given in terms of @xmath351 in eq . .
in this way we obtain @xmath352 . combining the last relation with @xmath346
, we find the metric of the original de sitter space in the new coordinates @xmath348 in the form @xmath353 @xmath354 is given by the version of eq . .
expressing then the factor @xmath355 by using eqs . , , and relations , we arrive at the de sitter metric in the accelerated coordinates in the form @xmath356 ^ 2 } \biggl(-\bigl(1-\frac{\ar^2}{\dsr^2}\bigr)\,\grad\at\formsq \\&\ ; + \bigl(1-\frac{\ar^2}{\dsr^2}\bigr)^{\!\!-1}\!\grad\ar\formsq + \ar^2\,\bigl(\grad\ath\formsq+\sin^2\ath\,\grad\ph\formsq\bigr)\biggr)\period\\ \end{split}\ ] ] here the accelerated coordinates @xmath357 are given in terms of static coordinates by the relation obtained by the procedure described above as follows:@xmath358 ^ 2}}\commae\\ \at=\stt\comma \tan\ath=\frac{\sqrt{1-{\ro^2}/{\dsr^2}}\;\str\sin\tht}{\str\cos\tht-\ro } \period \end{gathered}\ ] ] notice that the time coordinate of static and accelerated frames coincide . technically , this is easy to see from the first relation in eqs .
and eq . .
intuitively , this is evident since the uniformly accelerated particles are at rest in the static coordinates , as well as in the accelerated coordinates , the only difference being that they are located _ at the origin _ of the accelerated frame .
setting @xmath359 in eq . , we get @xmath360 , @xmath361 , as expected .
the static coordinates are centered on the poles @xmath362 , hence , on the _ _ un__accelerated worldlines .
the name _ accelerated _ coordinates is thus inspired by the fact that their origin _ is _ accelerated , and the value of this acceleration enters the form of the metric explicitly through the quantity @xmath363 .
the 2-dimensional conformal diagram of de sitter space with coordinate lines @xmath335 , @xmath364 of the accelerated frame is given in fig . [
fig : acc ] . for details , see the figure caption . here
let us just notice that the cosmological horizons are still described by @xmath365 .
infinite values of @xmath366 can , however , be encountered the conformal infinities @xmath324 are reached .
this depends on the angle @xmath325 .
indeed , @xmath367 corresponds to @xmath326 , whereas @xmath324 is given by @xmath327 , i.e. , @xmath368 ( cf .
relation of these two surfaces is best viewed in minkowski space @xmath186 .
we see that for @xmath369 , the conformal infinity @xmath370 ( @xmath371 ) lies ( ) the surface @xmath341 .
thus the infinity @xmath341 is just a coordinate singularity , which can be removed using , for example , the c - metric - like coordinate @xmath372 introduced below .
[ fig : accrth]a shows the cut @xmath373 located in the region of de sitter space where the killing vector @xmath374 is timelike ( @xmath375 ) ; @xmath76-direction is suppressed .
the cut is a _ spacelike _ sphere @xmath376 with homogeneous spherical metric .
the coordinate lines @xmath364 and @xmath332 are plotted , with two origins @xmath322 indicated : here the accelerated particles occur .
the coordinate @xmath377 grows from @xmath378 at the origins to the equator where @xmath379 . in fig .
[ fig : accrth]b the cut @xmath373 located in the regions where @xmath374 is spacelike ( @xmath380 ) is illustrated , again with @xmath76-direction suppressed . here
the cut is _ timelike _ with the geometry of three - dimensional de sitter space .
the coordinate lines @xmath364 and @xmath332 are also shown .
as we have just seen , the points with @xmath321 can be points in de sitter manifold
. it may thus be convenient to introduce the inverse of @xmath377 as a new coordinate .
also , we consider @xmath381 as a new coordinate , and make the time coordinate dimensionless .
we thus arrive at the : @xmath382 the metric becomes @xmath383 with the conformal factor @xmath384 given by @xmath385 this is de sitter spacetime in the : setting the mass and charge parameters , @xmath386 and @xmath387 , equal to zero in the the c - metric with a positive cosmological constant ( written in the form ( 2.8 ) of ref .
@xcite ) , and choosing the acceleration parameter equal to @xmath388 , we obtain the metric . in order to arrive naturally to the robinson - trautman form of the metric ,
notice that the coefficients in the metric become singular at @xmath389 , similarly as they do on the horizon of the schwarzschild spacetime in the standard schwarzschild coordinates .
analogously to that case , we choose a coordinate @xmath390 by @xmath391 similarly to the schwarzschild case again , we introduce a suitable _ null _ coordinate @xmath392 in terms of the radial and time coordinates @xmath393 and @xmath390 as follows : @xmath394 together with the conformal factor @xmath384 defined in eq . , we arrive at the de sitter metric in coordinates @xmath395 ( cf .
eq . ) which is very near to being in the robinson - trautman form
. however , there is a non - vanishing mixed metric coefficient at @xmath396 which is absent in the robinson - trautman metric .
such a term can be made to vanish by introducing a new angular coordinate @xmath397 by @xmath398 the de sitter metric then becomes @xmath399 where @xmath400 this is precisely the form of the robinson - trautman metric
see , e.g. , @xcite .
tracking back the transformations leading to the metric , the connection between the robinson - trautman coordinates and the static coordinates @xmath318 turns out to be not as complicated as our procedure might have indicated , in particular , for the radial coordinate .
we find a nice formula for @xmath401 , @xmath402 whereas the other two coordinates are simply expressed only in terms of accelerated coordinates @xmath403 : @xmath404 coordinates @xmath334 can then be obtained in terms of the original static coordinates by using eqs . .
the robinson - trautman coordinates with metric are centered on the accelerated particles . as with static or accelerated frames , we need two sets of such coordinates to cover both @xmath204 and @xmath200 . the relations to the static coordinates become , of course , much simpler if the particles are not accelerated , @xmath359 , and when both the robinson - trautman and static coordinates are centered on the pole @xmath117 : @xmath405 however , even robinson - trautman coordinates possess some very convenient features .
the radial coordinate @xmath384 is an affine parameter along null rays @xmath406 , normalized at the particle s worldline by the condition @xmath407{\rtr}\ , \mtrc_{\ds\mu\nu}\ , \tens{u}^\nu=-1\commae\ ] ] where @xmath408 is the particle s four - velocity .
these null rays form a diverging but nonshearing and nonrotating congruence of geodesics .
the null vector @xmath409 , tangent to the rays , is parallelly propagated along them . its divergence is given by @xmath410{\rtr}=2/\rtr}$ ] so that @xmath384 is both the affine parameter and the luminosity distance ( see , e.g. , @xcite ) . with robinson - trautman coordinates , one can also associate a null tetrad ( explicitly written down in the appendix , eq . ) which is parallelly transported along the null rays from the particle ( @xmath411 ) up to infinity ( @xmath412 ) .
the field at an event @xmath413 can be interpreted as @xmath414 of the sum of the retarded fields produced by particle @xmath73 at @xmath415 and particle @xmath74 at @xmath416 .
the field at @xmath417 can be interpreted as @xmath414 of the sum of the retarded and advanced effects from particle @xmath73 .
the affine parameter distances @xmath418 and @xmath419 are equal , the same being true for the distances @xmath420 and @xmath421 . ] owing to the boost symmetry of both the worldlines and de sitter space , an interesting feature arises , which is analogous to the situation in minkowski space .
consider a point @xmath417 in region @xmath116 ( fig .
[ fig : affpar ] ) .
there are two generators of the null cone with the origin at @xmath417 which cross the worldline @xmath73 at two points , @xmath422 and @xmath423 .
then the affine parameter distance @xmath424 is the same as @xmath425 .
( in order to go towards the past from @xmath423 to @xmath417 , the robinson - trautman coordinates built on the past null cones with origins on @xmath73 can easily be introduced . )
this is evident because @xmath417 lies on one orbit of the boost killing vector @xmath113 and a boost can be applied which leaves the worldline @xmath73 invariant but moves @xmath417 into event @xmath426 on the slice of time symmetry , @xmath226 ( also @xmath427 ) , where the particle is at rest .
then @xmath423 and @xmath422 move to the new points @xmath428 , @xmath429 , which are located symmetrically with respect to @xmath226 .
the equality of the affine parameter distances then follows from the symmetry immediately .
similarly , for an event @xmath413 in region @xmath114 one can show that the affine parameter distance @xmath430 is equal to the distance @xmath421 ( see fig .
[ fig : affpar ] ) .
the point @xmath413 lies on a boost orbit ( which now has a spatial character ) along which it can be brought , by an appropriate boost , to the point located symmetrically between the worldlines @xmath73 and @xmath74 ( lying so on the equator , @xmath276 ) .
the same consideration can , of course , be applied to an event in the region @xmath115showing that the affine distances along future - oriented null rays from an event to the particles are equal .
although the symmetries just described are common to the worldlines of uniformly accelerated particles in minkowski and de sitter spacetimes , an important difference exists . in minkowski space ,
the affine parameter distance along the null ray from an event on particle s worldline , such as @xmath415 , to an point @xmath413 is equal to the proper distance between @xmath431 and @xmath415 where @xmath431 is the _ orthogonal _ projection of @xmath413 onto the spacelike slice @xmath432 .
this is not the case in de sitter space if , as it appears natural , under an ` orthogonal projection ' we understand the projection of the observation point @xmath413 onto the spacelike slice @xmath433 containing @xmath415 performed along a timelike geodesic orthogonal to such a slice .
nevertheless , the proper distance @xmath434 between @xmath415 and @xmath431 is still related to the affine parameter distance @xmath384 by a simple expression : @xmath435 this relation can be derived as follows .
consider , without loss of generality , @xmath415 located at the turning point of the particle @xmath73 at @xmath436 .
the condition that the events @xmath437 and @xmath415 are connected by a null ray implies that the distance @xmath434 between @xmath415 and @xmath431 is the same as the time interval between @xmath431 and @xmath437 as measured by the metric of the conformally related static einstein universe .
since @xmath437 occurs at some time @xmath157 whereas @xmath438 and @xmath439 at @xmath246 ( i.e. , at static time @xmath436 ) , this time interval is equal to @xmath440 , cf .
the static radial coordinate @xmath108 of @xmath437 thus reads ( cf .
eqs . , )
@xmath441 the slice @xmath436 has a geometry of the 3-sphere of radius @xmath85 . using the standard law of cosines in spherical trigonometry for the sides of the triangle spanned by @xmath438 , @xmath439 , and the north pole
, we can eliminate @xmath442 . finally , employing eq .
, we obtain the result .
clearly , near the particle @xmath73 we have @xmath443 , and eq . then gives @xmath444 , as in minkowski space . in the following section we shall explore the character of the fields of the particles @xmath73 and @xmath74 .
we shall see that the affine parameter distance @xmath401 will play most important role , simplifying their description enormously .
namely , as we will see in section [ ssc : fcentonpart ] , eq .
, the affine parameter @xmath401 is identical to the factor @xmath445 which will be introduced in the following and will appear in all expressions for the fields .
in this section we wish to construct the scalar and electromagnetic fields of uniformly accelerated ( scalar and electric ) charges in de sitter universe . a general procedure , suitable in case of any not necessarily uniform acceleration would be to seek for appropriate green s functions .
alternatively , in particular for sources moving along uniformly accelerated worldlines , we can make use of the conformal relations between minkowski and de sitter spaces , and of the properties of scalar and electromagnetic fields under conformal mappings .
this method is advantageous not only for finding the fields in de sitter spacetime , but also for understanding their relationships to the known fields of corresponding sources in special relativity .
the only delicate issue is the fact that there are no conformal mappings between minkowski and de sitter space which are globally smooth .
we discussed , in section [ sc : coor ] , how various regions of one space can be mapped onto the regions of the other space . in ref .
@xcite we carefully treated the fields at the hypersurfaces where the conformal transformation fails to be regular . in order to obtain well - behaved fields
, one must continue analytically across such a hypersuface the field obtained in one region into the whole de sitter space . in section ii in ref .
@xcite , we also analyzed in detail the behavior of the scalar field wave equation with sources and of maxwell s equations with sources under ( general ) conformal transformations . in ref .
@xcite we primarily concentrated on the absence of purely retarded fields at the past infinity @xmath371 of de sitter spacetime in fact , in any spacetime in which @xmath371 is spacelike . in order to analyze this problem we also considered , in addition to monopole charges , more complicated sources like rigid and geodesic dipoles
; and we constructed some retarded solutions to show their patological features .
however , we confined ourselves to the sources the worldlines of which start and end at the poles ; we did not employ coordinates best suited for exhibiting the properties of the fields at future infinity @xmath370 , and the frames corresponding to cosmological models like flat ( @xmath446 ) or hyperbolic ( @xmath447 ) cosmological coordinates ; and we did not give the physical components of the fields . in the following
we shall find the fields and discuss their properties in various physically important coordinate systems , in particular those significant at @xmath370 or in a cosmological context . in the next section
, we also obtain the fields due to the uniformly accelerated scalar and electric charges starting at @xmath371 at @xmath276 ( ) . this , among others , will be important when we wish to regain the classical born fields in the limit @xmath448 .
we start by using the analysis of the conformal behavior of the fields and sources given in section ii in @xcite , and we also take over from @xcite the resulting forms of the fields due to the sources starting and ending at the poles of de sitter space , as described in standard coordinates . consider two uniformly accelerated point sources starting at @xmath371 ( i.e. , at @xmath225 , @xmath449 ) at the poles @xmath450 and @xmath451 ( fig .
[ fig : wldln_p ] ) .
their worldlines @xmath204 , @xmath200 are given by eqs .
( or ) in these standard ( rescaled ) coordinates , by eqs . and
in the flat and hyperbolic cosmological coordinates , and by eqs . in the static coordinates .
their simplest description is , of course , given by @xmath322 and @xmath411 in the accelerated and robinson - trautman coordinates since these frames are centered exactly on their worldlines . in section [ sc : accpart ]
we discussed physical velocities and other properties of these particles .
now , as noticed at the beginning of section [ sc : accpart ] , these _ two _ worldlines can be obtained by conformally mapping the worldline of _ one _ uniformly moving particle in minkowski space into de sitter space .
the fields of _ uniformly moving _ sources in minkowski space are just boosted coulomb fields . under a conformal rescaling of the metric , @xmath452 ,
the fields behave as follows : @xmath453 , @xmath454 ( see @xcite , section ii , where the behavior of the source terms is also analyzed ) . hence , the fields due to two _ uniformly accelerated _ sources in de sitter spacetime can be obtained by conformally transforming the boosted coulomb fields in minkowski spacetime .
employing the conformal mapping , we arrive at the following results ( note @xcite ) .
the scalar field is given by the expression @xmath455 where @xmath456^{\frac12 } \commae \end{split}\ ] ] or , written in the standard rescaled coordinates , @xmath457^{\frac12 } \!\period\ ] ] this field is produced by _ two _ identical charges of magnitude @xmath40 moving along worldlines @xmath204 and @xmath200 .
it is smooth everywhere outside the charges and it can be written as a symmetric combination of retarded and advanced effects from both charges ( cf .
( 6.6 ) in @xcite ) .
similarly to the scalar - field case , by using conformal technique the electromagnetic field produced by two uniformly accelerated charges moving along @xmath204 and @xmath200 can be obtained in the form @xmath458\commae \end{split}\ ] ] where @xmath459 is again given by eq . .
as in the scalar - field case , the field is smooth , non - vanishing in the whole de sitter spacetime and involving thus both retarded and advanced effects ( cf.section viia in @xcite ) .
however , an important difference between the scalar and electromagnetic case exists : the magnitude of the scalar charges is the same , whereas the electromagnetic charges producing the fields have _ opposite _ signs .
this is analogous to the situation in minkowski spacetime described in section [ sc : bornm ] ( see the discussion below eq . ) . at the root of
this fact appears to be cpt theorem cf .
@xcite for the analogous gravitational case where the masses of the particles uniformly accelerated in the opposite direction are the same . in de
sitter spacetime , as in any spacetime with compact spacelike sections , a simpler argument exists : the total charge in a compact space must vanish as a consequence of the gauss theorem @xcite . to gain a better physical insight into the electromagnetic fields , we shall introduce the orthonormal tetrad @xmath460 and the dual tetrad @xmath461 tied to each coordinate frame used , and we shall decompose the electromagnetic field @xmath462 into the electric and magnetic parts . such a decomposition , of course , depends on the choice of the tetrad .
for example , in the standard spherical coordinates @xmath104 the electromagnetic field ( 2-form ) @xmath462 can be written as @xmath463 and the electric and magnetic field spatial vectors are given in terms of their _ frame _ components as follows : @xmath464 in the present case of the standard spherical coordinates , using the explicit forms of the tetrad given in the appendix ( eqs . )
, we find @xmath465\commae\notag\\ & \emb_{\cfsp } = \frac{\emcharge}{4\pi}\ , \frac{\accl\dsr^2}{\kfact^3}\ ; \sinh\frac\spt\dsr\,\sin\spr\sin\tht\;\cbv{\ph } \period\label{emebsymsp}\end{aligned}\ ] ] in the appendix the orthonormal tetrads tied to the coordinate systems considered in this paper are all listed explicitly .
the only exception is the robinson - trautman coordinate system with one coordinate null and thus with a nondiagonal metric ; in this case the null tetrad is given in which the newman - penrose - type components are more telling .
the tetrad components of the electric intensity and the magnetic induction vectors are physically meaningful objects : they can be measured by observers who move with the four - velocities given by the timelike vector of the tetrad ( as , e.g. , @xmath466 for spherical cosmological observers ) , and are equipped with an orthonormal triad of the spacelike vectors ( e.g. , @xmath467 ) .
we first list the resulting electromagnetic field tensor and its electric and magnetic parts in the coordinate systems centered on the poles @xmath260 . the scalar field is always given by expression , the explicit form of the scalar factor @xmath459 changes according to the coordinates used .
since this factor enters all the electromagnetic quantities as well , we always give @xmath459 first and then write the electromagnetic field quantities . in the flat cosmological coordinates ( see eqs .
, ) we find : @xmath468^{\frac12 } \commae \end{split}\ ] ] @xmath469\commae \end{split}\ ] ] @xmath470\commae\\ \emb_{\cfcf } & = -\frac{\emcharge}{4\pi}\ , \frac{\dsr\sinh\acp}{\kfact^3}\,\frac{\mvr}{\dsr}\,\exp\frac\mvet\dsr\ , \sin\tht\;\cbv{\ph } \period \end{aligned}\ ] ] in the hyperbolic cosmological coordinates ( see eqs . , ) , the results are slightly lengthier : @xmath471^{\frac12 } \commae \end{split}\ ] ] @xmath472\!\commae \end{split}\ ] ] @xmath473\commae\notag\\ & \emb_{\cfch } = -\frac{\emcharge}{4\pi } \frac{\dsr}{\kfact^3 } \sinh\acp\cosh\frac\cht\dsr\sinh\frac\chr\dsr\sin\tht\;\cbv{\ph } \period\label{emebsymch}\end{aligned}\ ] ] much simpler expressions for the fields arise in the static coordinates ( see eqs .
we obtain @xmath474 @xmath475\commae \end{split}\ ] ] @xmath476\commae\notag\\ & \emb_{\cfst } = 0\period\label{emebsymtr}\end{aligned}\ ] ] since for practical calculations and for an understanding of the conformal relations between minkowski and de sitter spaces the rescaled coordinates are very useful , we also give the fields in these coordinates .
the rescaled coordinates are tied with the same orthonormal tetrad as non - rescaled ones , and they define the same splitting into electric and magnetic parts ( @xmath477 and @xmath478 are the same spatial vectors ) ; the functional dependence on the coordinates , however , is different . in the standard
rescaled ( conformally einstein ) coordinates ( see eqs . ) , which cover the whole de sitter spacetime including its conformal infinities globally , we get eq . for @xmath445 and @xmath479\commae \end{split}\ ] ]
@xmath480\commae\\ \emb_{\cfsp } & = -\frac{\emcharge}{4\pi}\ , \frac{\accl\dsr^2}{\kfact^3}\ ; \cot\tlt\sin\tlr\sin\tht\;\cbv{\ph } \commae \end{aligned}\ ] ] whereas in the flat rescaled cosmological coordinates , which cover globally the conformally related minkowski space ( see also fig .
[ fig : dsminkv ] ) , we arrive at @xmath481^{\frac12 } \commae\ ] ] @xmath482\commae \end{split}\ ] ] @xmath483\commae\\ \emb_{\cfcf } & = \frac{\emcharge}{4\pi}\ , \frac{\dsr\sinh\acp}{\kfact^3}\,\frac{\mvr}{\mvt}\ , \sin\tht\;\cbv{\ph } \period \end{aligned}\ ] ] in various contexts the electromagnetic field four - potential form @xmath484 , implying the field @xmath485 , may be needed . in the standard rescaled ( conformally einstein ) coordinates the potential reads @xmath486 from this expression the frame components can easily be obtained and the four - potential form can be transformed directly to any coordinate system of interest . the four - potential acquires a particularly simple form in static coordinates : @xmath487 inspecting now the expressions , we first notice few basic features of the fields . as a consequence of the axisymmetry , the azimuthal @xmath76 component of the electric field vanishes . on the other hand , only the azimuthal @xmath76 component of the magnetic field is non - zero . at the axis of symmetry , @xmath135 ,
the latitudinal @xmath488 component of the electric field and magnetic field vanish as @xmath489 .
the electric field points along the axis . in the classical born solution in minkowski space ,
both charges are , at any time , located symmetrically with respect to the equatorial plane @xmath490 .
consequently , the radial part of the electric field vanishes for @xmath490 ( cf .
eq . ) . in de
sitter spacetime the charges outgoing from the poles are , at all times , symmetrically located with respect to the sphere @xmath276 ( illustrated as the circle in fig .
[ fig : ds ] ) .
we thus expect the @xmath488 component of the electric field to vanish for @xmath276 .
this , indeed , follows from eq . .
this symmetry can be seen only in the standard spherical coordinates since the sphere @xmath276 is not covered by the hyperbolic cosmological coordinates and in the flat cosmological coordinates only one particle occurs .
another typical feature of the born solution in minkowski space is its time symmetry . as a consequence ,
the magnetic field vanishes at @xmath24 ( cf .
eq . ) . in the past , it was this fact which led some investigators , w. pauli @xcite among them , to the conclusion that there is `` no formation of a wave zone nor any corresponding radiation '' since @xmath491 at @xmath24 .
however , it is not at a spacelike hypersurface @xmath492 but at @xmath5 , which is reached by taking @xmath493 constant , @xmath494 , where the born field has typical radiative features , i.e. , @xmath495 ( see @xcite , @xcite , @xcite ) . in our generalized born solution , the time symmetry of the fields is clearly demonstrated in the global standard coordinates : under inversion @xmath496 the electric field in eq .
is invariant , whereas the magnetic field changes the sign ; @xmath497 at @xmath226 .
the field also exhibits radiative character when we approach @xmath5 in an appropriate way , as it is briefly indicated in @xcite . as mentioned in the introduction , a detailed analysis of the radiative properties of the generalized born field
will be given elsewhere .
the fields take the simplest form in the static coordinates , eq . .
in these coordinates the particles are at rest , and they both have a constant distance from the poles ; their world lines are the orbits of the killing vector @xmath113 of de sitter space .
the electric field is time independent , the magnetic field vanishes .
this is fully analogous to the born field in minkowski spacetime : it is static , and purely electric in the rindler coordinates , the time coordinate of which is aligned along the orbit of the boost killing vectors ( see , e.g. , @xcite ) . however , as we discussed in section [ sc : coor ] , the static coordinates cover only a of de sitter space . in the other half , the killing vector @xmath113 becomes spacelike .
it is in this non - static domain ( regions @xmath114 and @xmath115 in fig .
[ fig : affpar ] ) where we expect , in analogy with the results in minkowski spacetime , to find fields which have radiative properties .
@xmath113 is the killing vector also in the non - static regions , however , it is spacelike here , as it is typical for a boost killing vector in minkowski space .
the fields of uniformly accelerated charges in de sitter spacetime are invariant under the boosts along @xmath113 everywhere .
they are thus boost - rotation symmetric as the born fields in minkowski spacetime . in the cosmological coordinates , respectively , in their rescaled versions , the fields are , of course , time dependent . here
we expect the effects of the expansion / contraction of de sitter universe to be manifested . indeed , considering in any of the cosmological frames the spatial coordinates
fixed , and examining the fields along the timelike geodesics , we discover that the fields exponentially decay at large times , i.e. , as @xmath5 is approached .
more specifically , with the spherical coordinates @xmath248 fixed , the factor @xmath445 behaves as @xmath498 at large times @xmath271 , and hence , we obtain @xmath499 , @xmath500 , @xmath501 being constants .
the electric field thus becomes radial at large @xmath271 .
similarly , in flat cosmological coordinates we find @xmath502 , @xmath503 . in the hyperbolic cosmological coordinates the proper time @xmath172 appears instead of @xmath504 .
the rapid decay of the fields along timelike worldlines at large times is caused by the exponential expansion ( at large times ) of the spatial slices @xmath75 ( respectively @xmath505 ) .
although our fields are just test fields , their exponential decay is another manifestation of the : geodesic observers in spacetimes with @xmath70 see at large times these spacetimes to approach the de sitter universe exponentially fast the universe becomes ( see , e.g. , @xcite ) . clearly , as one approaches _ past _ infinity @xmath6 ( @xmath225 ) , the fields also decay exponentially .
it is interesting to notice the character of the field as it would be seen by the observers at rest with respect to the hyperbolic cosmological coordinates in the limit at which the particles the region covered by these observers across the horizon @xmath506 ( cf .
[ fig : dshyp ] ) , given in the hyperbolic coordinates by @xmath507 , @xmath508 .
as discussed in section [ sc : accpart ] , the observed velocity of the charges at this boundary is ( in the limit ) equal to the velocity of light . employing the transformation formulas , it is easy to see that at this boundary @xmath509 .
hence , the factor @xmath445 is finite here ( as it is evident from its scalar character and its finiteness in the global standard coordinates ) .
also , the radial part of the electric field remains finite .
however , @xmath510 diverges as @xmath511 here , indicating that the field has a character of an impulse , in fact , rather of an impulsive wave : indeed , eq . implies @xmath512 .
the situation appears to be analogous to the field of a static charge viewed from an inertial frame boosted to the velocity of light in minkowski spacetime ( see , e.g. , @xcite ) .
as expected , a remarkable simplification occurs when the fields are evaluated in the coordinates at the origin of which the charges are situated at all times . since the accelerated coordinates @xmath348 and the c - metric - like coordinates
are simply related by eqs .
, the discussion of the field properties is the same in both these frames .
namely , notice that both coordinate systems are tied with the same orthonormal tetrad , and they thus define the same splitting of the field into the electric and magnetic parts .
in these coordinates , we find the factor @xmath445 to read @xmath513 the scalar field is again given by @xmath514 , and the electromagnetic field also acquires now an extremely simple form : @xmath515 @xmath516 the magnetic field vanishes in the frame tied to the accelerated and c - metric coordinates , the electric field has precisely the coulomb form , with the factor @xmath445 playing the role of a distance . as signalized above already , the factor @xmath445 turns out to be the robinson - trautman radial coordinate ( see eq .
below ) , i.e. , the affine parameter distance allong null geodesics . the geometrical role of @xmath445
was elucidated in section [ ssc : rtcoor ] . considering a fixed point in de sitter universe and a light cone emanating from this point
, three typical situations can arise as illustrated in fig .
[ fig : affpar ] . for a point @xmath417 from the regions @xmath116 or @xmath118 ,
there are two null geodesics , one past - pointing , the other future - pointing ,
each of which crosses the worldline of _ the same _ particle , say @xmath73 ( in case of @xmath417 from @xmath116 ) , at points @xmath422 and @xmath423 ( see fig .
[ fig : affpar ] ) .
since @xmath445 is equal to the ( specific ) affine parameter distance which is the same from @xmath422 as from @xmath423 ( see section [ ssc : rtcoor ] ) , we can interpret the field as arising from purely retarded , respectively , advanced effects from @xmath422 , respectively @xmath423 ; or , equivalently , as a combination of retarded and advanced effects from these points . in the second situation , when the fixed point , say @xmath413 , is located ( in the region @xmath114 ) , there are two past - oriented null geodesics emanating from it which cross now _ both _ particles @xmath73 and @xmath74 at points @xmath415 and @xmath416 ( see fig .
[ fig : affpar ] ) . the field can be interpreted as arising from retarded effects only , either as a combination from both particles @xmath73 and @xmath74 , or as the retarded field from just one of them .
finally , for a point from the region @xmath115 the field can analogously be interpreted in terms of advanced effects . as we discussed in section
[ ssc : acoor ] and illustrated in detail in the appendix , the accelerated coordinates ( similarly as the static coordinates to which they go over for a vanishing acceleration ) are _ static _ ,
i.e. , the vector @xmath337 tangent to the orbits of the killing vector is timelike , only in the regions @xmath116 and @xmath118 ( cf . figs .
[ fig : acc ] , [ fig : affpar ] ) .
observers following the orbits of the killing vector are thus confined to the regions @xmath116 and @xmath118 , and they can not detect the fields in the region @xmath114 ( respectively @xmath115 ) .
nevertheless , notice that although the time coordinate @xmath517 diverges at the horizon @xmath120 , the radial coordinate @xmath366 is perfectly finite there , @xmath379 ( cf .
eq . with @xmath120 ) ,
and the field is meaningful in the region @xmath114 ( or @xmath115 ) as well . since here the roles of the coordinates @xmath366 and @xmath517 are interchanged , @xmath366 becoming a _ time coordinate _
, the field becomes _ time - dependent_. as mentioned above , we do not expect to find radiative properties in the regions @xmath116 and @xmath118 .
indeed , in accelerated coordinates the field is static coulomb field , with @xmath445 playing the role of a distance .
however , the radiative properties of the whole field in the wave zone in the region @xmath114 are not evident from the time - dependent , purely electric field in the accelerated coordinates with @xmath366 as a time coordinate .
it is worthwhile to recall that with finite sources in minkowski spacetime the field at _ any _ event is of a _ general _ algebraic type ; only asymptotically , at large distances , its features approach those of a ( @xmath518 , @xmath519 ) , if there is a radiation ( see , e.g. , @xcite ) . in case of
a non - null field , one can always introduce a frame in which the electric and magnetic fields are collinear , or , in the language of the newman - penrose formalism , to choose such a null tetrad @xmath520 , corresponding to the orthonormal tetrad , that the only non - vanishing null - tetrad component is @xmath521 ( see eqs . for the explicit expressions of the null tetrad and eqs . for the null - tetrad components of the electromagnetic field )
such a situation arises precisely for the null tetrad associated with the accelerated coordinates : the null - tetrad components are simply @xmath522 the vanishing of the other two null - tetrad components , @xmath523 and @xmath524 , has a deeper algebraic explanation : the null tetrad tied to the accelerated coordinates is special in the sense that it contains _ both _ principal null directions of the electromagnetic field . inspecting the form of the null tetrad constructed from the orthonormal tetrad
, we observe that both these principal null directions are tangent to the surfaces @xmath525 in the accelerated coordinates .
the radiative properties are well exhibited in the robinson - trautman coordinates . as we discussed in section
[ ssc : rtcoor ] , these coordinates are tied to the future null cones centered on the worldline of a particle .
we consider the null cones with vertices on the particle @xmath73 .
let us recall that the radial coordinate @xmath401 is the affine parameter along the generators of the null cones , each of which is given by @xmath526 fixed .
now , as mentioned above , it turns out that the _ factor @xmath445 is precisely equal to this affine parameter @xmath401 _ : @xmath527 the scalar field is then simply given by @xmath528 a remarkably nice form also acquires the electromagnetic field : @xmath529 the newman - penrose scalars are defined in terms of the null tetrad , which is parallelly propagated from the source to the along the rays @xmath530 .
they look as follows : @xmath531 now the radiative character of the field is transparent : the first term entering the peeling behavior , the scalar @xmath532 , decays indeed as @xmath533 , and it is non - vanishing for a non - zero acceleration @xmath250 . in the expressions and , the de sitter background is completely . the same form of the fields are obtained in case of uniformly accelerated charges in minkowski space if the coordinates built on the null cones emanating from the particles are employed .
a difference between both cases reveals itself only in the explicit dependence of the affine parameter @xmath401 on the coordinates of spacetime points .
finally , we turn to the fields from the particles symmetrically located with respect to the origin @xmath117 ( the ) of the standard spherical coordinates .
the particles are thus asymptotically at the equator , @xmath276 , at @xmath225 , and return back at @xmath270 with the opposite speeds ( fig .
[ fig : wldln_e ] ) .
their fields , of course , are intrinsically the same as those considered in the preceding section but only now they represent the _ direct _ generalization of the classical born solutions due to uniformly accelerated charges symmetrically located with respect to the origin of minkowski space
. we shall find the generalized born fields easily by using the transformation which we applied to obtain the worldlines of the particles born at the equator from those born at the poles .
the scalar field due to two equal scalar charges @xmath534 moving along the worldlines @xmath535and @xmath19 reads @xmath536 where the factor @xmath537 is determined by @xmath538^{\frac12}\commae \end{split}\ ] ] and the conformal factor @xmath539 is given by ( cf.eq . )
@xmath540 this factor is left in the explicit form here , in contrast to the preceding section , since it explicitly exhibits conformal relation of the scalar field under conformal mappings between de sitter space and minkowski space @xmath541 .
this relation will be used in the following to perform the limit from the born field in de sitter to the born field in minkowski spacetime .
the electromagnetic field produced by charge @xmath542 moving along the worldline @xmath535 and by symmetrically located charge @xmath543 moving along @xmath19 has the following form : @xmath544\commae \end{split}\ ] ] with factor @xmath537 given by . in the tetrad tied to the standard
spherical coordinates the electric and magnetic fields become @xmath545\commae\notag\\ & \emb_{\cfsp}^{\dsborn}=\frac{\emcharge}{4\pi } \frac{\accl\dsr^2}{\rfact^3}\ , \frac { \sinh\frac\spt\dsr\ , \sin\spr\,\sin\tht}{(1+\cosh\frac\spt\dsr\cos\spr)^3 } \;\cbv{\ph}\period\label{emebbdssp}\end{aligned}\ ] ] in the standard rescaled ( conformally einstein ) coordinates the expressions
slightly simplify : @xmath546^{\frac12 } } { \sin\tlt+\cos\tlr}\commae\label{rfacttl}\\ \omega_\minko = \frac{\cos\tlr+\sin\tlt}{\sin\tlt}\commae\end{gathered}\ ] ] @xmath547\commae \end{split}\ ] ] @xmath548\commae\\ \emb_{\cfce}^{\dsborn}&=-\frac{\emcharge}{4\pi } \frac{\accl\dsr^2\sin^2\tlt } { \rfact^3(\sin\tlt+\cos\tlr)^3 } \cos\tlt\sin\tlr\sin\tht\,\cbv{\ph}\period \end{aligned}\ ] ] the character of these fields was discussed in the preceding section for the particles @xmath73 and @xmath74 .
one must only rotate all the structures by @xmath156 in the @xmath549 direction ; hence , for example , the sphere of symmetry changes from @xmath276 to @xmath490 .
there is some interest in having the fields available also in the hyperbolic cosmological coordinates .
they cover only those regions of the fields in which we assume the radiative properties will be manifested .
the sources producing the fields are not covered by these coordinates ( cf.fig .
[ fig : wldln_e ] ) .
the fields in the hyperbolic cosmological coordinates look as follows : @xmath550^{\frac12}\commae \end{split}\label{rfactch}\\ \omega_\minko = 1+\cosh\frac\cht\dsr=2\cosh^2\frac\cht{2\dsr}\commae\end{gathered}\ ] ] @xmath551\!\commae \end{split}\ ] ] @xmath552 we shall use these results in the forthcoming paper on the radiative properties of the generalized born solution . here , finally , we wish to describe the limiting procedure which leads from the generalized born solutions directly to their counterparts in minkowski spacetime . for this purpose
it is natural to employ the conformally minkowski coordinates @xmath553 introduced in eq . , with the inverse transformation given in the appendix , eq . .
transforming the fields of the particles @xmath535 , @xmath19 from the conformally einstein coordinates to the conformally minkowski coordinates , we arrive at the following intriguing forms .
the scalar field is given by eq . where now the factors @xmath537 and @xmath539 are determined by @xmath554 notice that factor @xmath537 coincides with the expression in minkowski space .
the electromagnetic field reads @xmath555\commae \end{split}\ ] ] and the electric and magnetic parts of the field turn out to be @xmath556 \commae\\ \emb_{\cfmo}^\dsborn & = \frac{\emcharge}{4\pi}\ , \frac{1}{\rfact^3}\ , \frac{1}{\bo\omega_\minko^2}\ , \mot\,\mor\sin\tht\,\cbv{\ph } \period \end{aligned}\ ] ] to connect these fields with their counterparts in flat space , note that they are conformally related by the conformal transformation . under the conformal mapping
, the field @xmath557 must be multiplied by factor @xmath558 , which gives @xmath559 , and @xmath560 in remains unchanged .
the transformed fields then coincide with the classical born fields , , and . in order to see the limit for @xmath561
, we parametrize the sequence of de sitter spaces by @xmath0 , and identify them in terms of coordinates @xmath562 . as @xmath563 , eq
. implies @xmath564 , @xmath565 .
after the trivial rescaling of @xmath566 , @xmath567 by factor @xmath568 , the standard minkowski metric is obtained .
the limit of the scalar and electromagnetic fields and , in which @xmath569 is kept constant ( with @xmath570cf .
, leads precisely to the scalar and electromagnetic born fields and in flat space .
because of the rescaling of the coordinates by factor @xmath568 , we get the physical acceleration equal to @xmath571 , and the scalar field rescaled by @xmath572 . the explicit limiting procedure carrying
the generalized born fields in de sitter universe back into the classical born solution in minkowski space has thus been demonstrated .
since 1998 the observations of high - redshift supernovae indicate , with an increasing evidence , that we live in an accelerating universe with a positive cosmological constant ( for most recent observations see , e.g. , @xcite ) . vacuum energy seems to dominate in the universe and it is thus of interest to understand fundamental physics in the vacuum dominated de sitter spacetime . in the present work , we constructed the fields of uniformly accelerated charges in this universe .
they go over to the classical born fields in minkowski space in the limit of a vanishing cosmological constant . aside from some similarities found , the generalized fields provide the models showing how a positive cosmological constant implies essential differences from physics in flat spacetime . for example , advanced effects occur inevitably due to the spacelike character of the past infinity @xmath371 and its consequence the existence of the past particles horizons , respectively , of the of the particles worldlines . since de sitter spacetime , according to our present understanding , appears to be not only an appropriate basic model for studying future cosmological epochs , but it is commonly used also for exploring the inflationary era , various physical processes have been investigated in de sitter space from the perspective of the early universe , among them , the effects of quantum field theory . also in quantum contexts , however , problems arise from combining the causal structure of the full de sitter spacetime with the constraint equations ( see @xcite for a recent review ) . these problems
are associated with the in spacetimes with a spacelike @xmath371 .
we analyzed this issue in detail for the classical electromagnetic and scalar fields with sources in ref .
@xcite .
another intriguing implication of the rapid expansion of de sitter universe due to a positive cosmological constant is manifested in the exponential decay of the fields at large times .
we noticed this explicitly on the late - time behaviour of the fields due to accelerated charges . in the present paper we wished to give all details on the construction of the fields and on coordinate frames useful in understanding their various aspects , including their relation to their counterparts in flat spacetime .
we did not here analyze the radiative characteristics of the fields . in the introduction we indicated that radiative properties depend on the way in which a given point of infinity is approached .
this is briefly described at the end of our paper @xcite . in de sitter spacetime
it is not a priori clear , as it is in special relativity , how to define global physical quantities like energy or energy flux .
such issues connected with the question of radiation from will be considered in a future presentation .
the authors are grateful to the division of geometric analysis and gravitation , albert einstein institute , golm , for the kind hospitality . here
most of the work was done . j.b . also thanks the institute of astronomy , university of cambridge , for hospitality .
we also acknowledge support by grants gar 202/02/0735 of the czech republic and research project no . msm113200004 .
nine families of coordinate systems are here introduced , described analytically and illustrated graphically .
the corresponding forms of de sitter metric , orthonormal tetrads and interrelations between the systems are given .
all these systems are suitable for exhibiting various features of de sitter space ; two families are directly associated with uniformly accelerated particles . although the majority ( though not all ) of these coordinate systems undoubtedly appeared in literature in some form already , they are scattered and , as far as we know , not summarized as comprehensively as in the following . in the main text
we refer frequently to this appendix , but the appendix can be read independently .
we hope it can serve as a catalogue useful for analyzing various aspects of physics in de sitter universe . by a we mean the systems with the _ same _
coordinate lines ; e.g. , @xmath573 and @xmath574 where @xmath575 , @xmath576 , etc .
seven of our families have the same spherical angular coordinates @xmath577 , @xmath336 , accelerated and robinson - trautman coordinates mix three coordinates , only azimuthal coordinate @xmath336 remains unchanged .
the radial coordinates label directions pointing out from the pole and acquire only positive values .
however , transformations among coordinates take simpler forms if we allow radial coordinates to take on negative values as well .
this causes no problems if , denoting by @xmath579 and @xmath580 the prototypes of time and radial coordinates , we adopt the convention that the following two values of coordinates describe the same point : @xmath581 hence , intuitively we may consider a point with @xmath582 and @xmath168 fixed to lie on diametrically opposite side of the pole @xmath583 with respect to the point @xmath584 .
the orthonormal tetrad @xmath585 associated with a coordinate system is tangent to the coordinate lines and oriented ( with few exceptions ) in the directions of growing coordinates .
it is chosen in such a way that the external product @xmath586 of of the dual tetrad has always the same orientation .
since all forms of the metric contain the term the only component @xmath587 of the tetrad vector @xmath588 in coordinate frame @xmath589 is related to the @xmath577-component of @xmath590 as @xmath591 and we thus omit @xmath588 henceforth . in the standard newman - penrose null complex tetrad @xmath520 with only nonvanishing inner products
@xmath592 , @xmath593 , the electromagnetic field @xmath594 is represented by three complex components:@xmath595 the null tetrad can be specified directly ( as it will be done in the case of robinson - trautman coordinates in eq . ) , or it can be associated with any orthonormal tetrad , say @xmath596 , by relations @xmath597 here , @xmath598 and @xmath599 are timelike and spacelike unit vectors respectively , typically in a direction of and coordinate , and @xmath600 are spacelike unit vectors in angular directions , @xmath601 , @xmath602 . for each coordinate family
we give the diagram illustrating section @xmath603 with the radial coordinate taking on _ both _ positive and negative values .
the diagrams thus represent the history of the _ entire _ main circle of the spatial spherical section of de sitter universe .
the left and right edges of the diagrams represent the south pole and should be considered as identified ; the central vertical line describes the history of the north pole . recalling the meaning of the negative radial coordinate we could eliminate the left half of each of the diagrams by transforming it into the right one by replacements @xmath604 .
however , it is instructive to keep both halves for better understanding of the spatial topology of the sections .
all diagrams are compactified they are adapted to the standard rescaled coordinates @xmath128 ( see below ) .
the past and future conformal infinities are drawn as double lines .
the ranges of time and radial coordinates are shown , the orientation of coordinate labels indicates the directions of the growth of corresponding coordinates .
the values of the factors @xmath605 , @xmath606 , @xmath607 , @xmath608 , @xmath609 , @xmath610 and @xmath611 in various regions of de sitter space .
the factors are defined in eqs .
, , , , and , respectively .
the factor @xmath607 is used only in the expressions for static coordinates in the region where the killing vector is spacelike .
therefore , we indicated the values of @xmath607 only in those regions , although eq .
defines @xmath607 everywhere .
the factors @xmath607 , @xmath610 , and @xmath611 are defined only for any given section @xmath612 , but not as unique functions on the whole spacetime ( they are not symmetric with respect to the pole ) .
this is related to our convention using negative radial coordinates , cf . the text below eq .
] the first family consists of the or @xmath104 , and of the or @xmath613 ( where @xmath614 ) .
these coordinates cover de sitter spacetime globally .
they are associated with cosmological observers with homogeneous spatial sections of positive spatial curvature .
the coordinates are adjusted to the spherical symmetry of the spatial sections : @xmath549 , @xmath577 , and @xmath336 are standard angular coordinates .
the coordinate @xmath615 is a proper time along the worldlines of the cosmological observers given by @xmath616 .
the vector @xmath617 is a conformal killing vector which is everywhere timelike .
the rescaled coordinates @xmath613 can also be viewed as the standard coordinates of the conformally related einstein universe ; they cover smoothly both conformal infinities @xmath2 of de sitter spacetime .
@xmath621\\ \cbv{\spr}&=\frac1\dsr\cosh^{\!-1}\!\frac\spt\dsr\,\cv{\spr}=\frac1\dsr\sin\tlt\,\cv{\tlr}\commae\displaybreak[1]\label{apx : sphtetr}\\ % \notag\\[-9pt]\\[-6pt ] \cbv{\tht}&=\frac1\dsr\frac1{\cosh(\spt/\dsr)\sin\spr}\cv{\tht}=\frac1\dsr\frac{\sin\tlt}{\sin\tlr}\cv{\tht}\period\notag % \cbv{\ph}&=\frac1\dsr\frac1{\cosh(\spt/\dsr)\sin\spr\sin\tht}\cv{\ph}=\frac1\dsr\frac{\sin\tlt}{\sin\tlr\sin\tht}\cv{\ph}\period\notag\end{aligned}\ ] ] the first flat cosmological coordinate family consists of the @xmath635 and of the @xmath143 .
hypersurfaces @xmath636 are homogeneous flat spaces and coordinate lines @xmath637 are worldlines of cosmological observers orthogonal to these hypersurfaces .
they are geodesic with proper time @xmath149 , the vector @xmath638 is a conformal killing vector .
the coordinates cover de sitter spacetime smoothly , except for the past cosmological horizon , @xmath639 , of the north pole where @xmath640 .
the coordinates thus split into two coordinate patches and the horizon .
the domain above the horizon has a cosmological interpretation of an exponentially expanding flat three - space .
the rescaled coordinates can be viewed as inertial coordinates in the conformally related minkowski space @xmath328 , cf .
[ fig : dsminkv ] ; the domain above the horizon corresponds to the , @xmath641 , of @xmath328 , the domain below corresponds to the , @xmath642 .
@xmath643 \mvt=\signmv\,\dsr\exp\bigl(\signmv\frac\mvet\dsr\bigl)\commae\\ \signmv=\sign \mvt\period\label{apx : signmv}\end{gathered}\ ] ] the ranges of coordinates are @xmath644 with negative values of radial coordinate @xmath148 interpreted as described in eq . .
@xmath645\\ \cbv{\mvr}=\exp\frac{\signmv\mvet}{\dsr}\,\cv{\mvr}=\frac{\signmv\,\mvt}\dsr\,\cv{\mvr}\commae\\ \cbv{\tht}=-\frac\signmv\mvr\exp\frac{\signmv\mvet}{\dsr}\,\cv{\tht}=-\frac1\dsr\frac\mvt\mvr\,\cv{\tht}\period \end{gathered}\ ] ] the second flat cosmological coordinate family consists of the @xmath656 and of the @xmath657 . they can be built analogously to the flat coordinates introduced above , with north and south poles interchanged only .
they thus have similar properties : hypersurfaces @xmath658 are homogeneous flat three - spaces , coordinate lines @xmath659 are geodesics with proper time @xmath660 , and @xmath661 is a conformal killing vector .
the coordinates cover de sitter spacetime everywhere except the future cosmological horizon , @xmath662 , of the north pole ( i.e. , the past horizon of the south pole ) , and the rescaled coordinates can be viewed as inertial coordinates in the conformally related minkowski space @xmath663 . the @xmath553 can be understood as spherical coordinates in the conformally related minkowski space @xmath541 .
the coordinates do not cover de sitter spacetime globally they cover only a region around north pole , see fig .
[ fig : coorcmo ] .
the boundary of this region is given by the conformal infinity of the minkowski spacetime .
these coordinates are useful for studying the limit @xmath678 .
this family consists of the @xmath107 , @xmath108 , @xmath488 , @xmath76 and the .
the metric does not depend on time coordinate @xmath689the coordinates are associated with a killing vector . since the killing vector changes its character , the coordinates do not cover the spacetime smoothly .
we first describe the static coordinates in domains @xmath116 and @xmath118 , where the killing vector is timelike . in domain
@xmath116 the orbits of the killing vector ( corresponding to the worldlines of static observers ) start and end at the north pole , in domain @xmath118at the south pole .
they are orthogonal to slices @xmath433 , each of which consists of two hemispheres ( one in domain @xmath116 , the other in @xmath118 ) with homogeneous spherical 3-metric .
the distances between static observers ( measured within these slices ) do not change . since the static observers must overcome first the cosmological contraction and then the expansion , they move with a ( uniform ) acceleration . @xmath698 \brr % = \frac\dsr2\,\log\frac{\sin\tlt+\sin\tlr}{\sin\tlt-\sin\tlr}\ , = \frac\dsr2\,\log\bigl(\tan{\frac{\tlt+\tlr}2}\;\cot{\frac{\tlt-\tlr}2}\bigr ) \commae\\ \begin{aligned } & \exp\frac\brr\dsr=\sqrt{\frac{\sin\tlt+\sin\tlr}{\sin\tlt-\sin\tlr } } \commae\!\ ! & \sinh\frac\brr\dsr&=\frac{\sin\tlr}{\sqrt{\sin^2\tlt-\sin^2\tlr } } \commae\\ & \tanh\frac\brr\dsr=\frac{\sin\tlr}{\sin\tlt } \commae & \cosh\frac\brr\dsr&=\frac{\sin\tlt}{\sqrt{\sin^2\tlt-\sin^2\tlr } } \period \end{aligned}\notag\end{gathered}\ ] ] here we describe the and the from the preceding section in domains @xmath114 and @xmath118 where the killing vector is spacelike .
these domains extend up to infinity , namely , domain @xmath114 up to @xmath5 , domain @xmath115 up to @xmath6 .
the orbits of the killing vector start at the south pole and end at the north pole in @xmath114 , and they point in opposite direction in @xmath115 .
the motion along them could thus be characterized as a from one pole to the other .
the lorentzian hypersurfaces @xmath433 are homogeneous spaces with positive curvature , i.e. , 3-dimensional de sitter spacetimes .
the signature factors @xmath605 and @xmath607 are defined as @xmath704 -1\quad\text{in domain $ { \pdom}$}\commae \end{cases}\ ] ] and + @xmath705 the coordinates ranges are @xmath706 with negative values of radial coordinate @xmath108 and @xmath695 interpreted as described in eq . .
@xmath709 \brr = \frac\dsr2\,\log\bigl(-\tan{\frac{\tlt+\tlr}2}\;\cot{\frac{\tlt-\tlr}2}\bigr ) \commae\\ \begin{aligned } & \exp\frac\brr\dsr=\sqrt{\frac{\sin\tlr+\sin\tlt}{\sin\tlr-\sin\tlt } } \commae\!\ ! &
\abs{\sinh\frac\brr\dsr}&=\frac{\sin\tlt}{\sqrt{\sin^2\tlr-\sin^2\tlt } } \commae\\ & \tanh\frac\brr\dsr=\spcm\frac{\sin\tlt}{\sin\tlr } \commae & \cosh\frac\brr\dsr&=\frac{\abs{\sin\tlr}}{\sqrt{\sin^2\tlr-\sin^2\tlt } } \period \end{aligned}\notag\end{gathered}\ ] ] the third type of cosmological coordinates are the .
the hypersurfaces @xmath96 are homogeneous spaces with negative curvature , coordinate lines @xmath714 correspond to the worldlines of cosmological observers orthogonal to these slices , and the vector @xmath715 is a timelike conformal killing vector .
the coordinates cover spacetime only partially they can be introduced in two disconnected domains near the north pole , namely , in the past of the event @xmath246 , @xmath198 ( where @xmath716 ) , and in the future of this event ( where @xmath717 ) .
this family consists of the , and the ( @xmath393 being different from @xmath271 of the standard coordinates ) .
contrary to the previous cases the accelerated coordinates are centered on uniformly accelerate origins : @xmath322 corresponds to two worldlines with acceleration @xmath311 .
the transformation relations to the systems introduced above mix these three coordinates in general .
the accelerated coordinates are closely related to the static system .
their time coordinates coincide , @xmath724 , and coordinate lines @xmath725 are the same as those with @xmath726 .
both coordinate systems are identical for @xmath727 .
sections @xmath728 with @xmath375 have geometry of 2-sphere with parallels and meridians given by the coordinate lines of the static coordinates @xmath729 .
the lines of coordinates @xmath334 are the deformed version of static ones , their poles are shifted along meridian @xmath277 towards each other , cf .
[ fig : accrth ] . two conformal diagrams of sections @xmath730 ( @xmath731 on the right , @xmath732 on the left ) , adapted to the accelerated coordinates , are depicted in fig .
[ fig : cooracc ] .
the shape of the diagram varies with different values of @xmath333 ; indeed , the position of infinity is given by @xmath733 .
see also fig .
[ fig : acc ] for sections @xmath323 .
the c - metric - like coordinates rescale only the values of the accelerated coordinates and regularize the coordinate singularity @xmath341 .
de sitter metric in these coordinates is a zero - mass limit of the c - metric ( the metric describing accelerated black holes ; see , e.g. , @xcite ) .
finally , we use four parameters @xmath734 to parametrize the acceleration.they are related as : @xmath735 & \cosh\acp & \;=\ ; & \frac{\dsr}{\sqrt{\dsr^2 -\ro^2 } } & \;=\ ; & \frac{\bo^2+\dsr^2}{2\,\dsr\bo } & \;=\ ; & \sqrt{1+\accl^2\dsr^2 } \commae\notag\\[-12pt ] \\[-6pt ] & \tanh\acp & \;=\ ; & \qquad\!\frac\ro\dsr & \;=\ ; & \frac{\bo^2-\dsr^2}{\bo^2+\dsr^2 } & \;=\ ; & -\frac{\accl\dsr}{\sqrt{1+\accl^2\dsr^2 } } \commae\notag\\[-2pt ] & \exp\acp & \;=\ ; & \sqrt{\frac{\dsr+\ro}{\dsr -\ro } } & \;=\ ; & \quad\,\frac\bo\dsr & \;=\ ; & \sqrt{1+\accl^2\dsr^2 } - \accl\dsr \period\notag\end{aligned}\ ] ] @xmath736\!\!\commae\\ \mtrc\!=\!\rtr^2\mspace{-1mu}\biggl[-(\cy^2\!{-}1)\,\grad\ct\formsq{+}\frac1{\cy^2\!{-}1}\,\grad\cy\formsq { + } \frac1{1{-}\cx^2}\,\grad\cx\formsq{+}(1{-}\cx^2)\,\grad\ph\formsq\biggr]\!\!\commae\\[-12pt ] % \\
% \begin{split } % \mtrc=\rtr^2\,\biggl[&-(\cy^2 - 1)\,\grad\ct\formsq+\frac1{\cy^2 - 1}\,\grad\cy\formsq\\ % & + \frac1{1-\cx^2}\,\grad\cx\formsq+(1-\cx^2)\,\grad\ph\formsq\,\biggr]\commae % \end{split } \intertext{where } \sphmtrc{}'=\bigl(\grad\ath\formsq+\sin^2\ath\,\grad\ph\formsq\bigr)\commae\\[4pt ] \omega = \frac{\sqrt{1-{\ro^2}/{\dsr^2}}}{1+({\ar\ro}/{\dsr^2})\cos\ath } = \frac\rtr\ar=\frac{\rtr\,\cy}\dsr\commae\\ \rtr=\frac{\dsr}{\cy\cosh\acp-\cx\sinh\acp}=\omega\,\ar=\omega\,\frac\dsr\cy\period\end{gathered}\ ] ] @xmath737 @xmath739 \str\cos\tht=\frac{\ar\cos\ath+\ro}{1+({\ar\ro}/{\dsr^2})\cos\ath}\commae \\[3pt ] \str\sin\tht=\frac{\ar\sin\ath\sqrt{1-\frac{\ro^2}{\dsr^2 } } } { 1+({\ar\ro}/{\dsr^2})\cos\ath}\commae \\[3pt ] \frac{\str^2}{\dsr^2}= 1-\frac{\bigl(1-{\ar^2}/{\dsr^2}\bigr)\bigl(1-{\ro^2}/{\dsr^2}\bigr ) } { \bigl(1+({\ar\ro}/{\dsr^2})\cos\ath\bigr)^2}\commae \\[3pt ] \tan\tht=\frac{\ar\sin\ath\sqrt{1-\frac{\ro^2}{\dsr^2}}}{\ar\cos\ath+\ro}\period \end{gathered}\ ] ] the inverse relations have the same form with @xmath350 and @xmath348 interchanged only and @xmath182 replaced by @xmath740 .
@xmath743 \ar=\frac{\rtr\,\cosh\acp}{1-(\rtr/\dsr)\sinh\acp\cos\ath}\commae\\[6pt ] % = \frac{\rtr\cosh\acp}{1+\frac\rtr\dsr\,\sinh\acp\tanh\bigl(\rtp-\frac\rtu\dsr\sinh\acp\bigr)}\\ \abs{\,\tan\frac\ath2}=\exp\bigl(\rtp-\frac\rtu\dsr\sinh\acp\bigr)\commae\notag \\[12pt ] \ct=\frac\rtu\dsr\cosh\acp{- } \frac12\log\abs{\frac { \dsr{-}\rtr\,(\sinh\acp\cos\ath{+}\cosh\acp ) } { \dsr{-}\rtr\,(\sinh\acp\cos\ath{-}\cosh\acp)}}\!\commae\notag\\[3pt ] \cy=\frac\dsr{\rtr\cosh\acp}-\tanh\acp\cos\ath\commae\\[6pt ] % = \frac\dsr{\rtr\cosh\acp}+\tanh\acp\tanh\bigl(\rtp-\frac\rtu\dsr\sinh\acp\bigr)\\ \cx=\tanh\bigl(\rtp-\frac\rtu\dsr\sinh\acp\bigr)\commae\notag\end{gathered}\ ] ] where @xmath744 is given in terms of the robinson - trautman coordinates by the last equation .
if we introduce the spherical coordinates @xmath745 boosted with respect to the flat cosmological coordinates @xmath143 by a boost @xmath181 ( in the sense of minkowski space @xmath186 ) , we find that the accelerated coordinates @xmath320 are related to @xmath746 in exactly the same way as the static coordinates @xmath110 are related to the coordinates @xmath147 .
the boost @xmath747 , @xmath748 , @xmath749 , @xmath750 , rewritten in the spherical coordinates @xmath751 , @xmath752 , reads @xmath753 and relations analogous to eqs . and
are : @xmath754 in the ( or in their complex version @xmath759 ) , de sitter metric takes the standard robinson - trautman form . the coordinate @xmath760 is null , the coordinate @xmath401 is an affine parameter along coordinate lines @xmath761 .
these lines are null geodesics generating light cones with vertices at the origin @xmath411 .
the coordinates @xmath762 ( or @xmath763 ) are angular coordinates , however , they are not functions of the accelerated angular coordinates @xmath764 only ( cf . eq . ) . because @xmath764 have a clearer geometrical meaning , we list some formulas also in the mixed coordinate system @xmath765 .
the origin @xmath411 of the robinson - trautman coordinates is centered on the worldline of the uniformly accelerated observer moving with the acceleration @xmath766 .
the coordinates are thus closely related to the accelerated coordinates . the coordinates @xmath767 do not cover the whole spacetime smoothly .
they can be introduced smoothly in the future of the north pole , or in the past of the south pole . at the boundary of these two domains , @xmath768 .
since the robinson - trautman coordinates are closely related to the congruence of null geodesics , it is convenient to introduce the null tetrad which is parallelly transported along these geodesics @xmath761 : finally , we return back to the coordinate systems which employ standard coordinates @xmath577 , @xmath336 . time and radial coordinates
can be transformed into two null coordinates .
such null coordinates can be associated with most coordinate families introduced above .
coordinates @xmath779 are related to the standard coordinates ; @xmath780 and @xmath781 to the flat cosmological coordinates ; @xmath782 to the conformally minkowski ; and @xmath783 to the static coordinates . coordinate vectors @xmath784 , @xmath785 , etc .
, are the pairs of independent null vectors in the radial 2-slices @xmath786 .
we do not allow the radial coordinate to be negative in the definitions of null coordinates because this would interchange the meaning of @xmath787 and @xmath788 .
the null coordinates are thus drawn in the right half of fig .
[ fig : coornull ] only .
the relation of time and radial coordinates @xmath790 to the corresponding null coordinates @xmath791 is given by usual formulas : @xmath792 here @xmath793 stands for @xmath794 , @xmath795 , @xmath796 , @xmath797 , and @xmath798 respectively ; similarly with @xmath799 .
the coordinates @xmath781 , @xmath782 , and @xmath780 can be viewed as null coordinates in the conformally related minkowski spaces @xmath800 , @xmath541 , and @xmath186 ; these are shifted with respect to each other by @xmath801 in the direction of the conformally einstein time coordinate @xmath161 , or associated null coordinates : + @xmath802 the remaining coordinates @xmath783 are related to the conformally einstein null coordinates @xmath779 by the : @xmath803 here the sign factors @xmath804 and @xmath805 are given by @xmath806 relations , between null coordinates can also be rewritten as follows : @xmath807 @xmath808 @xmath809 @xmath810 @xmath811 + @xmath812 the same relations hold for coordinates @xmath788 .
, , , * * , ( ) , . the coordinates @xmath813 differ from the coordinates @xmath814 of @xcite ( where only case @xmath815 is considered ) by a simple rescaling @xmath816 , @xmath817 , and @xmath818 .
the conformal factor is then given by @xmath819 . | this paper deals thoroughly with the scalar and electromagnetic fields of uniformly accelerated charges in de sitter spacetime .
it gives details and makes various extensions of our physical review letter from 2002 .
the basic properties of the classical born solutions representing two uniformly accelerated charges in flat spacetime are first summarized .
the worldlines of uniformly accelerated particles in de sitter universe are defined and described in a number of coordinate frames , some of them being of cosmological significance , the other are tied naturally to the particles .
the scalar and electromagnetic fields due to the accelerated charges are constructed by using conformal relations between minkowski and de sitter space .
the properties of the generalized born solutions are analyzed and elucidated in various coordinate systems .
in particular , a limiting procedure is demonstrated which brings the cosmological born fields in de sitter space back to the classical born solutions in minkowski space . in an extensive appendix ,
which can be used independently of the main text , nine families of coordinate systems in de sitter spacetime are described analytically and illustrated graphically in a number of conformal diagrams .
-24pt j. bik and p. krtouj .
math .
phys . * 46 * , 102504 ( 2005 ) [ reformatted]fields of accelerated sources : born in de sitterj .
math .
phys . * 46 * , 102504 ( 2005 ) [ reformatted ] |
permeation of gas molecules through an amorphous polymer film is a problem of considerable scientific interest , as well as industrial significance ref1 . for example , the selective permeability of certain types of polymeric materials is often used for gas separation .
it is therefore of primary importance to understand how the diffusivity of different gases depend on various factors , such as the temperature , the size of the penetrant molecule , and the total amount and distribution of accessible free volume inside the polymer matrix @xcite . in a typical gas permeation experiment , a polymer film , formed by cooling the polymer from the rubbery state below the glass transition ,
is subjected to a pressure gradient across the film , so that gas molecules permeate the film .
after a time lag @xmath7 , the pressure on the low - pressure side starts to increase , and eventually , a stationary gas current across the membrane is established .
for very long times , the pressures on both sides of the film equilibrate ; this late stage is not of interest here .
the important parameters which are often used to describe the permation process are ( _ i _ ) the solubility @xmath8 , which gives the total amount of gas trapped in the membrane and ( _ ii _ ) the diffusion coefficient @xmath9 of a gas molecule inside the network . to determine the latter , one approximates the late - time dependence of the pressure by a straight line and obtains its intercept , @xmath10 , with the zero pressure axis .
the diffusion coefficient then follows from the relation @xmath11 , where @xmath12 is the thickness of the film .
the product of the solubility and diffusion coefficient is referred to as the permeability @xmath13 , and is a direct measure of the efficiency of the permeation process . from a slightly different point of view
, the gas permeation problem may be discussed in the language of ( electric ) resistor networks .
the pressure gradient set up across the membrane is analogous to a fixed voltage difference .
local configurational fluctuations inside the polymer matrix allow gas molecules to jump across energy barriers between voids , forming a connected set of `` resistors '' which control the flow of the gas current .
in fact , passage of ( electric ) current through a set of _ randomly _ distributed resistors is a problem that has been extensively studied in statistical physics @xcite . from this perspective
, gas diffusion through a polymer network is quite intriguing since the matrix of energy barriers and voids can hardly be considered random or uncorrelated .
it is therefore natural to explore the role of these correlations in the transport process .
the primary objective of this paper is to introduce a lattice - gas model with correlated , quenched energy barriers to study the transport of gas molecules through the polymer network . while we simplify the description significantly , by reducing it to two dimensions , and modeling polymers as simple non - interacting random walks , we nevertheless retain an essential feature of the original problem , namely , the effect of correlations induced by the polymer connectivity .
we find that the currents through this network depend very sensitively on ( _ i _ ) whether particle motion can be completely blocked by certain energy barriers ; and ( _ ii _ ) whether these spatially correlated barriers can percolate across the lattice .
our discussion draws on an earlier study @xcite in which we investigated the percolation properties of random walks of varying length .
the remainder of this paper is divided into three sections : section ii contains the model description , section iii discusses the monte carlo simulation results and section iv outlines our conclusions .
we consider a polymer network whose large scale structure is static over the time scale of the simulations .
this is a reasonable simplification since typical permeation experiments use polymers in the glassy state : such membranes offer better selectivity than those in a rubbery state .
small _ local _ rearrangements which let gas molecules pass from one void to another do occur and will be modeled via a distribution of energy barriers which the molecules need to overcome .
the polymer network is generated by placing @xmath14 non - interacting random walks of length @xmath15 ( referred to as `` @xmath15-mers '' ) on a two - dimensional square lattice with @xmath16 sites .
each _ bond _ of the lattice can carry one or more segments ( i.e. , steps of the random walk ) , with an average segment density per bond of @xmath17 .
the gas molecules are modeled as hard - core particles which occupy the _ _ cells _ _ of the lattice , and no more than one particle is allowed per cell . while modeling polymers as non - interacting ( rather than self - avoiding ) walks is almost certainly an oversimplification , we can offer two motivations for such a simple description .
first , real membranes are produced by rapid cooling of a rubbery melt ; hence , typical polymer configurations are `` frozen '' high - temperature states for which interactions are less important .
second , when discretizing a real network , it is natural to choose the lattice spacing to be comparable to the polymer persistence length . in this case
, a polymer `` segment '' contains many monomers . since the persistence length is typically much larger than the size of a monomer unit , we have in effect adopted a coarse - grained description . on this length scale , self - avoidance and other complications may be neglected .
each polymer segment constitutes an energy barrier in the path of a gas molecule .
the dynamics of the molecules is modeled as activated hopping over these barriers .
once a particle - hole pair , separated by such a barrier , is chosen , the probability for them to exchange is @xmath3 , where @xmath4 is the number of segments in the barrier and @xmath5 is just @xmath18 , with @xmath19 being an energy scale and @xmath20 being the inverse temperature . as an alternate view , one may consider the particle motion as resulting from the short time scale segmental motion of the polymer .
if the probability of displacement of a single segment is @xmath5 , then @xmath3 would be the probability of simultaneously displacing all @xmath4 segments . to model the pressure gradient ,
the cells in the first row ( @xmath21 ) of the lattice are kept filled at all times . at the beginning of the simulation ( time @xmath22 ) , the remainder of the lattice is empty . at times
@xmath23 , each particle randomly selects a direction of motion with probability @xmath24 , and attempts to cross the barrier between its originating and target cells . provided
the target cell is empty , the particle will be moved with probability @xmath3 .
after every particle in the system has attempted a move ( on average ) , @xmath25 is incremented by one monte carlo step ( mcs ) .
when a particle arrives in a cell within the last row ( @xmath26 ) , it is removed instantaneously from the lattice .
clearly , this sets up a density gradient across the system which drives a particle current in the @xmath27-direction . in the simulations
, we measure the following quantities : 1 . the number of particles which have entered the system up to time @xmath25 , denoted by @xmath28 2 . the number of particles which have exited the system up to time @xmath25 , denoted by @xmath29 .
the mean first passage time , @xmath7 , for the arrival of the _ first _ particle at the far boundary @xmath26 . at any time @xmath25 , therefore , the number of particles absorbed by the polymer network is given by @xmath30 . at sufficiently late times , the system approaches a steady state , which is characterized by a time - independent solubility @xmath31and a constant particle current , or permeability , @xmath32 , defined via @xmath33to give a reference point , the current through an empty lattice would be @xmath34 .
the mean first - passage time provides information about the typical path length travelled by a particle , and how quickly it hops over the barriers encountered along that path .
the numerical simulations were carried out for various values of the segment density @xmath35 and polymer length @xmath15 .
we choose a lattice with @xmath36 sites for the simulations , which was also the largest lattice simulated in @xcite .
we choose polymer lengths in powers of @xmath37 : @xmath38 with @xmath39 .
the typical span of even the longest polymer ( @xmath40 ) is much smaller than the lattice dimension , so that a single polymer only very rarely extends across the whole lattice .
our choices for the segment density @xmath35 were guided by the phase diagram of the @xmath6 system which we will summarize in the next section .
the smallest value of @xmath41 was @xmath42 and the largest was @xmath43 .
intermediate values ( e.g. @xmath44 ) are controlled by @xmath14 , the ( integer ) number of polymers of length @xmath45 placed in the system . for different @xmath15 s , we selected @xmath14 s to be within less than a percent of @xmath46 .
so far , we considered just two values of the bond - crossing probability : @xmath47 and @xmath48 . due to computational limitations , we generated between @xmath49 and @xmath50 different polymer configurations for each set of parameters @xmath51 .
@xmath52 and @xmath53 are collected for @xmath54 mcs for each configuration and then averaged .
let us comment briefly on statistical errors , which can be estimated by comparing currents for different configurations at the same @xmath55 .
they are smallest ( below @xmath56 ) for the shortest polymers and @xmath48 . as @xmath57 increases
, the errors also become larger , reaching about @xmath58 for @xmath59 at this @xmath5 . for @xmath47 and especially @xmath60 ,
finite - size effects allow for both blocked and open configurations at the same @xmath55 , making it difficult to assign error bars reliably . a conservative estimate would give errors as large as the currents themselves , so that our conclusions can only be preliminary until better statistics are available .
when the bond - crossing probability @xmath5 vanishes ( or , equivalently , the zero temperature limit ) , bonds carrying one or more segments become completely impenetrable . as a consequence , the gas current must vanish if the occupied bonds form a connected path ( a `` spanning cluster '' ) transverse to the density gradient .
the problem , therefore , reduces to bond percolation of @xmath14 @xmath15-mers on a square lattice .
we studied this problem in @xcite , and summarize only the salient results here . with the segment density ( per bond ) given by @xmath61 ,
the fraction of occupied bonds , @xmath62 , approaches a well - defined value in the thermodynamic limit . in our simulations
, we found that @xmath63 decreases with @xmath15 for fixed @xmath35 and increases with @xmath35 for fixed @xmath15 .
while the latter behavior is obvious , the former needs some explanation . to make it more intuitive ,
let us consider two situations where the density @xmath35 is kept fixed , but the polymer lengths ( and hence , their numbers ) are different .
configurations with few , long polymers are likely to exhibit numerous bonds which are multiply occupied since a simple random walk retraces any given step with probability @xmath24 .
in contrast , such multiple bond crossings are far less likely for short polymers , even at the same @xmath35 , resulting in a larger @xmath63 . to rephrase
, single segments leave far fewer bonds unoccupied ( `` free '' ) than longer polymers . with a qualitative understanding of @xmath62 ,
it is natural to ask how the percolation threshold , @xmath64 , is crossed as one increases @xmath41 at given @xmath15 . from simulations , we find that @xmath64 is a _ monotonically decreasing _ function of @xmath15 , i.e. , when @xmath15 is large , a smaller number of occupied bonds already suffices to ensure the presence of a spanning cluster .
we believe that this is primarily due to long polymers forming extended cigar - shaped objects @xcite which percolate more easily @xcite than single occupied bonds or short polymers .
clearly , two consequences of polymerization compete with one another : the elongated shape of the polymers tends to lower , while the presence of multiple occupancies tends to increase , the percolation threshold . due to this competition ,
the critical segment density per bond,@xmath65 , is a _ non - monotonic _ function of @xmath15 , shown in fig . 1 for a @xmath36 square lattice .
we see that @xmath66 increases from @xmath67 for @xmath68 to @xmath69 for @xmath70 but decreases beyond that to @xmath71 ( @xmath72 ) and @xmath73 for @xmath74 . for shorter polymers ,
@xmath66 is dominated by the effect of multiple occupancies ( which increase with @xmath15 ) , but for longer ones , their tendency to form extended objects which percolate more easily , takes over .
when @xmath15 becomes comparable to @xmath12 , finite - size effects become noticeable , causing @xmath66 to increase again : @xmath75 for @xmath76 . here , multiple occupancies dominate again , this time due to large sections of _ different _ polymers overlapping one another .
not surprisingly , this effect is more pronounced on smaller lattices , making it difficult to extrapolate the behavior of @xmath77 .
whatever the eventual resolution of these finite - size effects , in the thermodynamic limit @xmath66 plays the role of a phase boundary : a gas current can flow only if @xmath78 @xmath66 . our earlier findings
@xcite provide an answer to a simple `` yes / no '' question : will a gas current flow or not ?
we now turn to a much more detailed question : if a current does flow , what are its quantitative characteristics ? by measuring currents
, we are probing not only the mere existence of open channels , but also their number and their length : how many distinct paths are there for a gas molecule to travel from one edge of the system to the other , and how many cells will it visit along a given path ? again , we use a combination of simulations and simple intuitive arguments to explore the answers .
we begin with systems in the low temperature limit @xmath47 , i.e. , occupied bonds can not be crossed at all , and then consider the effects of softening this constraint by having non - zero @xmath5 .
our numerical results for the steady - state particle current @xmath79 are presented in figs . 2 and 3 .
in fig . 2 , we show @xmath0 plotted vs. polymer length @xmath15 , for fixed @xmath80 and @xmath47 . according to the results summarized in the previous section , this value of @xmath35 is well below the percolation density @xmath66 for all @xmath15 considered here ; hence the current is non - zero for these @xmath15 .
in contrast , the inset shows @xmath0 vs. @xmath15 at a higher density , @xmath81 , which is within a few percent of the critical @xmath66 .
in fact , for @xmath82 , @xmath35 lies slightly _ above _ the critical @xmath83 , and consequently , no current should flow for this @xmath15 .
as one might anticipate , the absolute values for the currents are much smaller ( by at least a factor of @xmath84 ) for the larger @xmath35 , since more bonds are blocked in that case . on a more quantitative level
, we observe very different behaviors at the two densities . for @xmath80
, @xmath0 decreases with @xmath15 , for all but the smallest @xmath15 .
in contrast , at @xmath81 , the current first increases , then drops off again to a minimum at @xmath74 , after which it recovers .
thus , polymerization has very different effects on the current , at low and high segment density .
we now attempt to interpret these findings , in the light of our percolation study summarized above .
considering , first , @xmath80 , we note that this value is well below @xmath66 for all @xmath15 considered . in other words , a typical configuration exhibits only finite clusters of occupied bonds , allowing for numerous current - carrying channels with a length of @xmath85 . polymerizing single segments ( @xmath68 ) into short polymers ( say , @xmath86 )
frees some occupied bonds ( by increasing multiple occupancies ) but has little effect on the number or the length of the conducting channels .
thus , the current for @xmath70 is , at best , slightly enhanced over the current at @xmath68 .
however , as the polymers become longer , they also extend further , forcing particles to flow around them .
moreover , longer polymers begin to form `` loops '' which enclose a fair fraction of free , but now inaccessible , bonds .
hence , having a few long polymers in the system can increase the effective length of the conducting channels significantly and thus decrease the current , as illustrated by fig . 2 .
turning to the inset of fig . 2 , the picture changes drastically .
this density ( @xmath81 ) is within a few percent of the percolation threshold , @xmath66 ; if any open channels exist , they are rare and fractal
. even minor rearrangements of occupied bonds suffice to open a new channel , or block an existing one . turning to specific @xmath15 s , we recall that @xmath87 is slightly above the percolation threshold for _ single _ segments : @xmath88 .
hence , for @xmath68 no current should flow , at least in the thermodynamic limit .
however , finite - size effects wash out a sharp percolation threshold , so that a very small current is observed : a few of the simulated polymer configurations did not percolate .
if single segments are tied together into longer polymers with , say @xmath70 , the number of occupied bonds is lowered significantly , since higher occupancies will be generated . as a result
, a much larger segment density is required for percolation : @xmath89 . translated into currents
, this implies that the @xmath70 system will carry a significant current at @xmath87 , as borne out by the data shown in the inset .
yet , the tendency of polymers to form extended objects competes with this trend .
we believe that the minimum , observed at @xmath74 , is genuine ( even though it lies well within the error bars ) due to the fact that this system is extremely close to the percolation threshold : @xmath90 .
if it were not for finite - size effects , we would expect the current to vanish here . to summarize , @xmath91 is dominated by the behavior of @xmath92 : @xmath0 vanishes if @xmath93 , and otherwise roughly follows @xmath66 : the distance to the percolation threshold @xmath66 controls the number of open channels which , in turn , control the current .
of course , there is no need to simulate systems with densities @xmath93 since the currents are expected to vanish .
let us now compare these findings to currents measured in systems with finite energy barriers . specifically
, we choose @xmath48 for the rate with which a singly occupied bond is crossed by a gas molecule .
for a bond occupied by @xmath4 segments , this rate drops to @xmath3 .
3 shows the corresponding currents for @xmath80 and @xmath94 ( inset ) .
both data sets show currents @xmath95 which appear to increase gently with @xmath15 . in terms of absolute values , @xmath96 is larger than @xmath97 because fewer bonds are occupied at the lower density .
of course , @xmath96 is also larger than @xmath98 : even though both cases have ( on average ) the same number of occupied bonds , these are blocked completely in the latter , but not the former case . for @xmath48
, a singly occupied bond allows gas molecules to cross in one out of @xmath84 attempts . in order to block efficiently ,
a bond should carry , say , three or more segments , so that the rate of crossing is reduced to @xmath99 or less . while such `` highly occupied ''
bonds exist , of course , they are extremely unlikely to form large connected clusters , even for the largest @xmath45 : they constitute less than about @xmath100 of bonds for @xmath80 , and less than about @xmath101 for @xmath94 . as a result , there are no extended objects that might impede the flow of particles ; instead , the currents are dominated by the number of completely free ( @xmath102 ) and singly occupied ( @xmath103 ) bonds . unlike the @xmath47 case , domains of unoccupied bonds inside polymer loops can now be accessed and contribute to the currents . this effect is particularly stark for systems above the percolation threshold : e.g. , at density @xmath94 with @xmath6 , all currents would vanish .
regarding the small dip observed in the @xmath80 data , we have no convincing explanation to offer at this stage . with statistical errors of about @xmath100
, it seems hardly significant .
yet , it is conceivable that the @xmath104 data reflect , in some very loose sense , the `` sum '' of a residual current flowing only across connected free bonds ( given by @xmath105 ) , plus a contribution carried by the remaining ( isolated ) free bonds and singly occupied ones . since the latter is likely to increase monotonically with @xmath15 , while @xmath98 is monotonically decreasing
, one might interpret the dip in @xmath96 as being the result of `` adding '' the two trends .
better data are needed before these questions can be settled satisfactorily . in very general terms
, the current @xmath0 may be thought of as dependent on two major factors : how many particles are present inside the matrix at any given time , and how fast they move across barriers , on average . in order to quantify these
, we measured the particle ` solubility ' @xmath108 in the steady state and the mean ` first passage time ' @xmath7 , which is the time it takes for the first particle to emerge out of the network at its far end . if we adopt a simplistic view and imagine that the ( barrier - hopping ) motion of a single particle inside the polymer network is simply diffusion with an effective diffusion coefficient @xmath9 , then the first passage time @xmath109 provides a measure of @xmath9 : longer time intervals @xmath7 imply lower @xmath110 .
we find that the solubility @xmath108 roughly trails the current data : higher currents are associated with larger @xmath108 , and vice versa . given that higher currents imply more open channels ( which can then be filled with additional particles ) and fewer ( if any ) inaccessible domains , this seems plausible .
our data for the mean first passage time in @xmath111 systems are unfortunately very noisy , so that no reliable conclusions can be drawn . for the @xmath48 systems
, @xmath7 decreases with @xmath15 , indicating that particles encounter fewer obstacles on their way through the network .
again , this broadly mirrors the current data .
in this paper , we studied a simple model of gas transport through a polymer membrane .
the membrane is modeled as a network of non - interacting random walks , placed on the bonds of a square lattice in two dimensions .
the network is characterized by the total segment density , @xmath35 , and the length , @xmath15 , of the random walks .
the gas molecules are modeled as hard - core particles performing activated hopping from one lattice cell to a nearest neighbor . when they cross a bond carrying @xmath4 segments , they encounter an energy barrier @xmath3 , with @xmath112 . at one end of the lattice ,
the particle density is held at unity ; at the other end , particles are immediately removed .
these boundary conditions drive particles across the network . to quantify the transport properties of the network , we focus primarily on the gas current , @xmath32 , in the steady state .
we explore in particular how polymerization at fixed segment density ( i.e. , joining a given number of segments into polymers of varying length ) affects the currents .
we also measure the total number of particles stored in the network , as well as the mean first passage time for the first particle to arrive at the far end of the lattice .
it is immediately obvious that this problem will be most interesting , and most difficult , in the low temperature regime , when the polymers form significant obstacles .
considering therefore small values of @xmath5 , including @xmath113 , our observations suggest that the transport properties of such a polymer network are affected by a competition of two main features .
( _ i _ ) when the total segment density is fixed , the fraction of _ multiply _ occupied bonds increases since random walks frequently `` retrace '' their last step . as a result , polymerization generates a larger number of completely _ free _
( @xmath102 ) bonds which facilitate particle motion .
( _ ii _ ) as the polymers increase in length , their radius of gyration also increases , as @xmath114 .
the polymers form extended , cigar - shaped objects which may enclose domains of unoccupied bonds and block particles very efficiently , especially for @xmath115 . as a result , for @xmath6 ,
the current is controlled by the underlying polymer percolation problem : it is strictly zero when the polymers form a connected cluster of occupied bonds transverse to the density gradient .
exploring the associated percolation threshold , @xmath66 , as a function of @xmath15 , we find that increasing the fraction of free bonds tends to shift it to higher densities while creating extended objects and `` loops '' pushes it towards lower densities ; hence , @xmath66 is actually non - monotonic in @xmath45 .
not surprisingly , the behavior of the currents is very subtle : for @xmath116 ( well below the percolation threshold ) , the current decreases with @xmath15 , as larger polymers block more efficiently ; for @xmath81 , the current is non - monotonic , reflecting the behavior of @xmath66 .
for nonzero @xmath5 , specifically @xmath48 , the picture changes significantly .
the finite extent of the polymers becomes far less important when particles can cross occupied bonds with nonzero probability . as a result
, the dominant factor is now the number of bonds which allow for high particle throughput , namely , @xmath102 bonds .
since their number increases with @xmath15 , the currents generally increase
. of course , the picture might become more complex again when the segment density becomes excessively large , generating a large number of `` highly occupied '' bonds . to summarize , the transport properties of a random walk network are highly nontrivial .
clearly , we have explored only a small domain of the full parameter space , and our statistical errors are still large . moreover , to reflect real polymers , more sophisticated models should surely be invoked , and more surprises may well be in store . * acknowledgements . *
it is a pleasure to acknowledge fruitful discussions with e. marand and h. hilhorst . over the years
, both bs and rkpz have enjoyed many hours with l. schfer who taught us much of what we know about polymers .
this work was partially supported by nsf dmr-0088451 and dmr-0414122 .
gm thanks the max planck gesellschaft for financial support and computational resources .
s. kirkpatrick , rev .
* 45 * , 574 ( 1973 ) ; i. webman , phys . rev .
lett . * 47 * , 1496 ( 1981 ) ; a. b. harris and t. c. lubensky , phys .
e * 35 * , 6964 ( 1987 ) ; o. stenull , h. k. janssen and k. oerding , phys . rev .
e * 59 * , 4919 ( 1999 ) .
diehl and e. eisenriegler , j. phys .
a * 22 * , l87 ( 1989 ) ; f. fougre and j. desbois , j. phys .
a * 26 * , 7253 ( 1993 ) ; l. schfer , _ excluded volume effects in polymer solutions , as explained by the renormalization group_. springer , berlin ( 1999 ) . | we study the diffusion of gas molecules through a two - dimensional network of polymers with the help of monte carlo simulations .
the polymers are modeled as _ non - interacting _ random walks on the bonds of a two - dimensional square lattice , while the gas particles occupy the lattice cells . when a particle attempts to jump to a nearest - neighbor empty cell , it has to overcome an energy barrier which is determined by the number of polymer segments on the bond separating the two cells .
we investigate the gas current @xmath0 as a function of the mean segment density @xmath1 , the polymer length @xmath2 and the probability @xmath3 for hopping across @xmath4 segments . whereas @xmath0 decreases monotonically with @xmath1 for fixed @xmath2 , its behavior for fixed @xmath1 and increasing @xmath2 depends strongly on @xmath5 . for small , non - zero @xmath5
, @xmath0 appears to increase slowly with @xmath2 .
in contrast , for @xmath6 , it is dominated by the underlying percolation problem and can be non - monotonic .
we provide heuristic arguments to put these interesting phenomena into context . |
in non equilibrium statistical mechanics , the boltzmann transport equation ( bte ) is a widely used tool for describing the transport properties of a classical thermodynamic system .
since this equation is applicable mainly to low density systems for which binary collisions dominate , it gives , for example , the most suitable way of analysing the composition and dynamics of the upper atmosphere of the earth . in this paper ,
the properties of the solution of this equation are explored in a manner to be of value primarily in the study of thermosphere .
the boltzmann equation , when applied to a gas mixture like the one in consideration in upper atmosphere , leads to many difficult situations .
the general interactions are varied and complex .
moreover , highly non - equilibrium conditions are found , for example , at high latitudes in the thermosphere , where the ions move at extremely high speeds compared to the neutral atoms due to the presence of magnetospheric electric fields .
another example can be the case of exobase ( lower boundary of the exosphere ) where the extremely sparse population fails to give a sufficiently high rate of collisions to justify the applicability of the collision term of the bte . in this case
, one would need a way to find whether or not the equation is valid . in order to address situations like these
, the ordinary collision integrals used to simplify the bte , fail to provide an advantageous insight in the situation due to the insurmountable complexity of the integrals .
the methods previously equipped to attempt at solving the bte include the eigenvalue approach of chapman and cowling @xcite , simplification of the collision term by chapman and enskog @xcite and orthogonal polynomial method by grad @xcite . of these , the method used by grad which primarily exploits the tensor hermite polynomials to obtain a series solution to the transport equation , presents explorable prospects .
+ in this paper , we study the tensor hermite polynomials and their properties to understand their applicability in areas other than dynamics of upper atmosphere involving applications in computational fluid dynamics , study of magneto plasma in tokomak for fusion research , dynamics in the core of a nuclear reactor , etc . the basic work on tensor hermite polynomials was done by grad in 1949 @xcite introducing the need for them .
he introduced them as the basis for complete set of orthonormal polynomials in n variables in tensorial form .
as mentioned in his work , this set of scalar orthonormal polynomials often lacks symmetry and an advantage is gained by expressing these polynomials in tensor invariant notation .
other than grad , work on these tensor polynomials has been done by viehland @xcite to solve the bte for trace amounts of ions in dilute gases and by knio @xcite to study the uncertainty propagation in computational fluid dynamics .
+ in this paper , three major aspects of the tensor hermite polynomials are studied . in the first one ,
the effects of scaling are examined .
this directly provides a mathematical criterion governing the ambient conditions that must be satisfied for justifying the applicability of the collision integral term of the bte in gas dynamical procedures .
the second studies the change in hermite polynomials under a translation of axes .
the third one inspects the polynomials under rotation of axes from particle velocity coordinates to the centre - of - mass and relative velocity coordinates .
this results in tensors of order 6 which are required to evaluate the collision integrals for a two component system of low density gas mixture .
the tensorial hermite polynomials can be written iteratively as ( [ app : a3 ] ) : @xmath0,\ ] ] where the permutation operator @xmath1 is explained in [ app : a2 ] . hence , the first few polynomials are & _ 0 = 1 , [ eq:2 ] + & _ 1 = 2 , [ eq:3 ] + & _ 2 = 4_i_j - 2_ij , [ eq:4 ] + & _ 3 = 8_i_j_k - 4(_i_jk + _ j_ki + _ k_ij ) . [ eq:5 ] the entities stated in bold are tensors .
the rank of these entities are denoted by the number of underlines below it or by the numerical subscript .
these polynomials differ from the tensor hermite polynomials introduced by grad @xcite described in [ app : a1 ] and are related to the above definition .
+ these polynomials are orthonormal and symmetric .
hence , any scalar function , for example , the distribution function of the boltzmann transport equation @xmath2 , can be expanded in terms of these hermite polynomials @xmath3 provided the integral @xmath4 here , @xmath5 is the dimensionless velocity vector such that : @xmath6 where @xmath7 is the particle velocity , @xmath8 is its mass , @xmath9 is the boltzmann constant and @xmath10 is the thermodynamic temperature . in equation [ eq:7 ] , @xmath11 is the weight , which , in this context is the normalised maxwellian velocity distribution function displaced by an amount @xmath12 . @xmath13 the @xmath14 s in equation [ eq:6 ]
are called the expansion coefficients .
they can be obtained by exploiting the orthonormality of the hermite polynomials described by @xmath15 [ eq:9 ] + the notation @xmath16 represents the scalar product of the tensors @xmath17 and @xmath18 ( [ app : a4 ] ) .
in this section we explore the possibility of conversion from one type of functional dependence to another where the other one is scaled with respect to @xmath5 .
consider the relation , @xmath19 where @xmath20 and @xmath21 are constants .
henceforth , summation over repeated indices is implied unless stated otherwise . considering the general functional dependence of the form , @xmath22 can be written in terms of @xmath5 as @xcite , @xmath23 [ eq:11 ] + if the same function has to be equivalently representable in terms of @xmath24 , assuming the same functional dependence , through the expression : @xmath25 [ eq:12 ] + then the series @xmath26 should be convergent . here ,
@xmath27 and @xmath28 are constants . from the above expressions ,
the two series are related as : @xmath29 } ( \boldsymbol{a}_n , \boldsymbol{h}_n),\ ] ] [ eq:13 ] + and the series in @xmath24 will converge if the following integral _ -^ = ( f_0/f_0)^2 _ -^ , exists . from the orthogonality of the hermite polynomials , we can write @xmath30 are functions of quadratic combinations of @xmath31 s . for a finite value of @xmath32 ,
the expression in equation [ eq:14 ] will be finite since @xmath33 depends on finite powers of @xmath34 .
hence , the existence of the integral depends only on that of the integral of the exponent , i.e. , @xmath35}\enspace dz } < \infty.\ ] ] [ eq:15 ] + this happens when @xmath36 has a negative coefficient in the exponent . by substituting for @xmath37 from equation ( [ eq : eq10 ] )
, we get the condition as @xmath38 .
+ from the expression of @xmath34 in equation ( [ eq : eq8 ] ) , we can see that @xmath39 where @xmath40 is the temperature of the species with the scaled velocities and @xmath10 is the temperature of particles with unscaled velocity .
+ hence the interchange of variables can be done if @xmath41 .
+ in studies of ion - neutral mixture in upper atmosphere , where @xmath42 ( @xmath43 is the ion temperature and @xmath44 is the neutral atom temperature ) , interconversions to and from a temperature @xmath10 is possible when it falls in the range : @xmath45 [ eq:16 ] + this also implies that for the boltzmann equation to hold , with the collision term , the ion temperature must be less than four times the neutral atom temperature .
translation can be made from one frame of reference to another .
there are generally three frames of reference : the inertial , the one moving with the entire gas , and the one moving with a particular component species .
let two translations be defined as : & _ 0 = - _ 00 , [ eq:17 ] + & _ r = - _ a , [ eq:18 ] where @xmath46 is an arbitrary constant and @xmath47 is the average value of @xmath5 . + the hermite polynomials under both the translations are inter - related through : & _ n^0 = s_(n,3)[_p=0^n ( _ p^n ) _
p^r [ 2(_a-_00)]^t(n - p ) ] , [ eq : eq19 ] + & _ p^r = s_(p,3)[_n=0^p ( _ n^p ) _ n^0 [ 2(_00 - _ a)]^t(p - n ) ] .
[ eq : eq20 ] in the above expressions , @xmath48 and @xmath49 represent the hermite polynomials in terms of @xmath20 and @xmath50 defined as per equation [ eq:17 ] and [ eq:18 ] respectively . here
, @xmath51 is the usual binomial coefficient .
these expressions can be obtained by the principle of mathematical induction and using equation ( 1 ) as explained in [ app : a7 ] .
it should be realised that translation destroys the property of orthogonality . in order to evaluate the collision term of boltzmann equation for the gas flow problem
, one has to consider the collision integral and the collision cross section depends on the relative velocity of the collision partners @xcite .
more often than not , this relative velocity can have a direction different from that of the species velocity . in that case ,
a rotation of axes makes the calculation a lot less cumbersome .
+ considering for example , @xmath52 and @xmath53 as the space vectors and @xmath54 and @xmath55 as the corresponding distribution functions for two species , the axes can be rotated to : & c_r = ( m_s v_rs + m_s v_rs)/ , [ eq:21 ] + & g_r = ( v_rs - v_rs ) , [ eq:22 ] + & v_rs = z_rs , [ eq:23 ] + & v_rs = z_rs , [ eq:24 ] + & = m_sm_s/(m_s + m_s ) , [ eq:25 ] where @xmath56 and @xmath57 are the rotated coordinates , @xmath58 and @xmath59 are the molecular masses of the two species , @xmath60 is the boltzmann constant , @xmath10 is the temperature , @xmath61 and @xmath62 are the species velocities and @xmath63 is the reduced mass . + the rotation matrix can be expressed as : & r_1;ssgc = r_1;gcs s = y & y + y & -y , + & y^2 = / m_s , y^2 = / m_s. and the corresponding relations can be expressed as : @xmath64 [ eq:26 ] + here , the vectors that are getting transformed are of order 6 , i.e. , two components corresponding to the two species each of order 3 .
the two components can be represented by upper and lower indices .
hence , we can also define the mixed hermite polynomials : & _ 0 = 1 , [ eq:27 ] + & _ 1 = ( _ 1a _ 1b)^t , [ eq:28 ] + & _ * h*__n+1 = s_(n+1,3)^(n+1,2 ) ( _ * h*__n _ * h*__1 - 2n ^u _ * h*__n-1 ) , [ eq:29 ] where @xmath65 and @xmath66 represents the permutation operator over the indices of both the components .
+ therefore , using @xmath67 and @xmath68 , we can define two sets of hermite polynomials .
it is observed that the distribution function is the same in rotated and in original axes frame ( [ app : a7 ] ) : [ eq : eq30 ] f_ss = f_rsf_rs = f_cg .
the description of tensor hermite polynomials and a study of their behavior under scaling , translation and rotation of axes gives us an insight on their applicability in various situations . since the assumptions in the process were minimal , they provide quite an accurate and generalised computational procedure .
this procedure ensures that with the inclusion of each additional tensorial polynomial in the truncated series , the discrepancy between the actual distribution function and the approximated one decreases monotonically .
the investigation of these polynomials under scaling gives a condition on temperature which must be satisfied for the collision term of the bte to be applicable .
the polynomials when treated under rotation , provide a simplified way of evaluation .
it should be noted that because of their orthonormality and symmetry , they can be advantageously exploited in simplifying systems other than the one given by boltzmann equation . + as further research in this area , we plan to work on exploring the properties of the expansion coefficients under various transformation of variables .
we also plan to look at the tensorial hermite polynomials in spherical polar coordinates which can be useful for certain systems .
one of the authors ( pm ) wishes to thank tata consultancy services , mumbai for funding this project and sarthak bagaria for helping at various stages of the project .
general formulae @xcite : & _ iz_j = _ ij , + & _ j^n_i^n = _ ij^n , + & _ ( n ) = ^n , + & _ ( n+1 ) = ( z_i - _ i)_(n ) , [ eq : eqa5 ] + & _ i_(n ) = _ i_(n-1 ) , [ eq : eqa6 ] + & _ ( n ) = ( - ) ^n . , + & . where @xmath70 is the hermite polynomial with indices @xmath71 and @xmath72 represents the normalized weight : @xmath73 + and the following holds @xcite & _
i= -_i , + & _ i ( ) = .
@xmath74 is a second order unit tensor , i.e. , the identity matrix and @xmath75 represents sum of all terms in which @xmath76 is attached with @xmath74 as explained in reference @xcite .
the relation between physicist s polynomial used in this paper and mentioned in the thesis of sengupta @xcite and the probabilist s polynomial used by grad @xcite is as follows : & _ n(z ) = 2^_(n)(z ) , + & _ ( n)(z ) = 2 ^ -_n(z/ ) . where @xmath69 is the probabilist s hermite polynomial .
now substituting for @xmath78 in equation [ eq : eqa6 ] gives : @xmath79 replacing @xmath80 by @xmath34 , we get : & _ i_n = 2_i_n-1 . similarly , in equation [ eq : eqa5 ] , substituting for the probabilist s polynomial
, we get : @xmath81 , % \\&or \frac{1}{\sqrt{2}}h_{n+1 } = \frac{1}{\sqrt{2}}[2x_i - \nabla_i]h_n , \\&or \quad \boldsymbol{h}_{n+1 } = ( 2x_i - \nabla_i)\boldsymbol{h}_n . \end{split}\ ] ]
the permutation operator @xmath82 acts on an argument and replaces the argument by a sum of all possible permutations over n indices for tensors in m - dimensional space @xcite .
+ for example , & s_(2,3 ) ( ) = _ ij + _ ji , + & s_(3,3)(x ) = x_i_jk + x_i_kj + x_j_ki + x_j_ik + x_k_ij + x_k_ji .
@xmath83 \\ & = \ ; s_{(n+1,3)}[\boldsymbol{h}_{n}\boldsymbol{h}_{1 } ] - s_{(n+1,3)}[2n\boldsymbol{h}_{n-1}\underline{\underline{i } } ] , \\ & = \ ; [ 2x_{i_1}s_{(n,3)}(\boldsymbol{h}_n ) + 2x_{i_2}s_{(n,3)}(\boldsymbol{h}_n ) + 2x_{i_3}s_{(n,3)}(\boldsymbol{h}_n ) + \dots ] \\ & \ ; \ ; \;- 2n[\delta_{i_1i_2}s_{(n-1,3)}(\boldsymbol{h}_{n-1 } ) + \delta_{i_1i_3}s_{(n-1,3)}(\boldsymbol{h}_{n-1 } ) + \dots \\ & \qquad \qquad \delta_{i_2i_1}s_{(n-1,3)}(\boldsymbol{h}_{n-1 } ) + \delta_{i_2i_3}s_{(n-1,3)}(\boldsymbol{h}_{n-1 } ) + \dots \\ & \qquad \qquad \dots \\ & \qquad \qquad \dots \qquad + \qquad \dots \qquad + \delta_{i_{n+1}i_n}s_{(n-1,3)}(\boldsymbol{h}_{n-1 } ) ] .
\end{split}\ ] ] here , @xmath1 the permutation operator explained in the previous section .
+ since the @xmath84 are the same for any permutations of the indices , we have @xmath85 therefore the above expression is : ( putting @xmath86 ) @xmath87 = & n![2x_{i_1}\boldsymbol{h}_n + 2x_{i_2}\boldsymbol{h}_n + \dots ] \\&-2n(n-1)![\delta_{i_1i_2}(\boldsymbol{h}_{n-1 } ) + \delta_{i_1i_3}(\boldsymbol{h}_{n-1 } ) + \dots \\ & \qquad \dots \\ & \qquad \dots \qquad + \qquad \dots \qquad + \delta_{i_{n+1}i_n}(\boldsymbol{h}_{n-1 } ) ] , \\= & n![2x_{i_1}\boldsymbol{h}_n + 2x_{i_2}\boldsymbol{h}_n + \dots ] \\&-2n![\boldsymbol{\delta}_{i_1}(\boldsymbol{h}_{n-1 } ) + \boldsymbol{\delta}_{i_2}(\boldsymbol{h}_{n-1 } ) + \dots + \boldsymbol{\delta}_{i_{n+1}}(\boldsymbol{h}_{n-1 } ) ] .
\end{split}\ ] ] where the @xmath75 s are obtained from the expression in equation [ eq : eqa6 ] .
@xmath87 = & n![2x_{i_1}\boldsymbol{h}_n + 2x_{i_2}\boldsymbol{h}_n + \dots ] \\&-n![\nabla_{i_1}\boldsymbol{h}_n + \nabla_{i_2}\boldsymbol{h}_n + \dots + \nabla_{i_{n+1}}\boldsymbol{h}_n ] , \\= & n![(2x_{i_1}-\nabla_{i_1})\boldsymbol{h}_n + ( 2x_{i_2}-\nabla_{i_2})\boldsymbol{h}_n + \quad \\ & \dots \quad + ( 2x_{i_{n+1}}-\nabla_{i_{n+1}})\boldsymbol{h}_n ] .
\end{split}\ ] ] using the relation : @xmath88 we get : @xmath87 = & n![\boldsymbol{h}_{n+1 } + \boldsymbol{h}_{n+1 } + \quad \dots \quad + \boldsymbol{h}_{n+1 } ] , % \\= & n!(n+1)h_{n+1 } , \\= & ( n+1)!\boldsymbol{h}_{n+1}. \end{split}\ ] ]
the scalar inner product of two tensors * a * and * b * of _ n_-th order is defined as @xcite : @xmath89 where n is the number of dimensions in the space under consideration .
+ for example , in 3 dimensions , for two second order tensors ( matrices ) , the scalar inner product will be given as : & ( ^(2).^(2 ) ) = a_11b_11 + a_12b_12 + a_13b_13 + & + a_21b_21 + a_22b_22 + a_23b_23 + & + a_31b_31 + a_32b_32 + a_33b_33
@xmath90 where , @xmath21 is a tensor of rank @xmath91 and order 6 and @xmath92 is the corresponding set of rotated mixed tensor expansion coefficients .
@xmath93 is the n - th rank rotation tensor .
this proof can be done by principle of mathematical induction . substituting for the first few values of @xmath32 or @xmath94
, we observe that the expressions hold true . assuming that the expression @xmath95^{t(n - p)}],\ ] ] holds true for @xmath32 then proving that this implies it also holds true for @xmath96 shows that this is true in general . from equation ( 1 ) , we have @xmath97 since the @xmath98 are symmetric in indices , @xmath99 and @xmath100 cancel . + subsituting for the series expression of @xmath48 @xmath101}^{t(n - p)}2(\underline{z}-\underline{z}_{00 } ) - 2n\sum_{p=0}^{n-1 } ( _ p^{n-1})\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p-1)}\underline{\underline{i } } , \\ & = \sum_{p=0}^n ( _ p^n)\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p)}2(\underline{z}-\underline{z}_{00 } ) \\ & \qquad \qquad - 2\sum_{p=0}^{n-1 } ( _ { p+1}^{n})(p+1)\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p-1)}\underline{\underline{i } } , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n)}2(\underline{z}-\underline{z}_{00 } ) + \sum_{p=1}^n ( _ p^n)\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p)}2(\underline{z}-\underline{z}_{00 } ) \\ & \qquad \qquad - 2\sum_{p=1}^{n } ( _ { p}^{n})(p)\boldsymbol{h}^r_{p-1}{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p)}\underline{\underline{i } } , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1 ) } + { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n ) } \boldsymbol{h}^r_{1 } + \sum_{p=1}^n ( _ p^n){[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p ) } \\ & \qquad \qquad \quad [ \boldsymbol{h}_p^r\boldsymbol{h}^r_{1 } + \boldsymbol{h}_p^r2(\underline{z}_a-\underline{z}_{00 } ) - 2p\boldsymbol{h}^r_{p-1 } \underline{\underline{i } } ] , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1 ) } + { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n ) } \boldsymbol{h}^r_{1 } + \sum_{p=1}^n ( _ p^n)\boldsymbol{h}^r_{p+1}{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p ) } \\ & \qquad \qquad + \sum_{p=1}^n ( _ p^n)\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p+1 ) } , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1 ) } + { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n ) } \boldsymbol{h}^r_{1 } + \sum_{p=1}^n ( _ p^n)\boldsymbol{h}^r_{p+1}{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p ) } \\ & \qquad \qquad + \sum_{p=0}^{n-1 } ( _ { p+1}^n)\boldsymbol{h}^r_{p+1}{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n - p ) } , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1 ) } + ( n+1){[2(\underline{z}_a -\underline{z}_{00})]}^{t(n ) } \boldsymbol{h}^r_{1 } \\ & \qquad \qquad + \sum_{p=1}^{n-1 } ( _ { p+1}^{n+1})\boldsymbol{h}^r_{p+1}{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1-(p+1 ) ) } + \boldsymbol{h}^r_{n+1 } , \\ & = { [ 2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1 ) } + ( n+1){[2(\underline{z}_a -\underline{z}_{00})]}^{t(n ) } \boldsymbol{h}^r_{1 } \\ & \qquad \qquad + \sum_{p=0}^{n-2 } ( _ { p}^{n+1})\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1-p ) } + \boldsymbol{h}^r_{n+1 } , \\ & = \sum_{p=0}^{n+1 } ( _ { p}^{n+1})\boldsymbol{h}_p^r{[2(\underline{z}_a -\underline{z}_{00})]}^{t(n+1-p)}. % \end{dmath } % \end{split}\end{aligned}\ ] ] | a description of orthogonal tensor hermite polynomials in 3-d is presented .
these polynomials , as introduced by grad in 1949 @xcite , can be used to obtain a series solution to the boltzmann transport equation .
the properties that are explored are scaling , translation and rotation .
order 6 hermite tensors are studied while obtaining the rotation relations . from the scaling of the independent variables of particle velocities , a criterion on temperature
is obtained which implies that the equation can be applied to binary gas mixtures only if the temperature of the hotter constituent is less than four times that of the cooler one .
this criterion and other properties of the tensor hermite polynomials obtained in this paper can be used to study gas dynamics in the thermosphere . |
recently we proposed @xcite to use a lateral interface between two regions of the two - dimensional ( 2d ) electron gas with different strengths of the bychkov - rashba @xcite spin - orbit ( so ) interaction as a spin - polarizing element for the purposes of spintronics . dattadas1990,wolf2001
the lateral interface introduces the space - varying so interaction which leads to spin - dependent refraction of spin carriers passing through the interface.@xcite consequently , an electron beam with a nonzero angle of incidence after passing through the interface splits into beams with different spin polarizations propagating in different directions ( see fig .
[ fig : refraction ] for the scattering at the lateral interface between the regions with and without so interaction ) . using an interface with an inhomogeneous so interaction as a principal element of spin - based devices we outlined the schemes for a spin filter , spin guide , and a spin current switch ( spin transistor ) .
this program promises to build spintronics devices avoiding magnetic materials that are not conventional for the semiconductor industry .
it can be realized in the gated heterostructures with a sufficiently strong bychkov - rashba so interaction @xcite by manipulating the gates . @xcite
the effect of the separation of the trajectories of electrons with different spin polarizations ( chiralities ) influenced by the space - varying so interaction has the same grounds as the double refraction ( birefringence ) of light in uniaxial crystals exploited in optical devices for the polarization of light .
the separation of the trajectories of carriers of different spin has been observed recently@xcite in the case of a homogeneous so interaction as a result of action of a perpendicular magnetic field .
the separation of the trajectories after reflection at a lateral potential barrier in the presence of the so interaction has been observed in ref . .
two facts that can be useful for the purposes of the spintronics have been found in ref . in the analysis of the spin - dependent scattering of electrons incident on a lateral interface with a so interaction varying in the direction normal to the interface .
first , there exists an interval of outgoing angles @xmath0 where only electrons with a definite spin chirality can penetrate ( see fig .
fig : refraction ) .
if it is possible to collect electrons from this interval , one will have an ideal spin filter .
second , electrons of this chirality exhibit a total internal reflection for an angle of incidence in an interval @xmath1 where @xmath2 is a critical angle of the total internal reflection .
it is clear from these observations that the electrons propagating at small angles to the lateral interface are most sensitive to the variation of the magnitude of the spin - orbit interaction , and therefore such electrons are most appropriate for spin control and manipulation .
hence one has to find a way to create ( and collect ) flows of electrons of high intensity that are almost tangential to the interface . in this paper
we suggest using the diffuse emission@xcite as a possible solution of this task .
we show that making one of the two regions connected by the lateral interface to be diffusive is an effective way to achieve a flat angular distribution of particles emitted from the diffusive region into the clean one .
the effect of flattening of the angular distribution of the emitted electrons is the robust property of the diffuse emission which holds despite the presence of a spin - dependent reflection at the interface . due to the flatness of the angular distribution a substantial portion of electrons
propagates at small angles to the interface and these electrons are suitable for the spin filtration and subsequent spin manipulation . in fig .
fig : diffuse - filter a scheme of a device is presented which operates as a spin filter with a high level of spin - polarization of a filtered current .
the paper is organized as follows . in section
[ sec : spinorbit ] we present the results@xcite of the analysis of scattering of electrons at the lateral interface between the two regions with different magnitudes of the bychkov - rashba term . in section [ sec : kinetics ] we analyze the transport of the 2d electrons in the presence of the coulomb interaction near the interface between the diffusive and ballistic regions . in section sec : milne we reconsider the problem of the diffuse emission ( milne problem@xcite ) for an arbitrary dimension . in section [ sec : spindependentemission ]
we generalize the technique developed for the problem of the diffuse emission to the case of the spin - dependent scattering at the lateral interface .
we use a semiclassical approach with electrons moving along the classically allowed trajectories when the spin of electrons is the only element treated quantum - mechanically . for this purpose
, we analyze the transport near the interface in terms of the spin density matrix . in the summary a scheme of a spin filter device that creates two beams of electrons of a very high level of spin polarization is presented .
consider a two - dimensional electron gas confined in the @xmath3 plane by a potential well in the semiconductor heterostructure .
generally , the potential well has the shape of an asymmetric triangle , and , consequently , there is a direction of asymmetry , @xmath4 , perpendicular to the electron gas plane .
this leads to the appearance of the specific spin - orbit interaction term@xcite in the hamiltonian , @xmath5 . we will consider the case when the parameter @xmath6 varies along the @xmath7 direction .
the direction of @xmath4 is chosen as @xmath8 generally , the hamiltonian has the form : @xmath9 .
\label{eq : hamiltonian}\end{aligned}\ ] ] here @xmath10 describes the varying bottom of the conduction band which may be controlled by gates .
the current operator corresponding to this hamiltonian contains a spin term : @xmath11 the presence of spin in the current operator implies that in the process of scattering at the lateral interface with varying @xmath6 the continuity conditions for the wave function involves the spin degrees of freedom of the electrons .
this makes the electron scattering at the interface to be spin - dependent . to diagonalize the hamiltonian with the bychkov - rashba term in the regions of constant @xmath6 one has to choose the axis of the spin quantization along the direction @xmath12 .
then for the two chiralities ( referred to below as @xmath13 and @xmath14 modes ) the dispersion relations are given by @xmath15correspondingly , the momenta of the waves of a given energy @xmath16 involved in the scattering at the interface between the two regions with different @xmath17 are determined as follows : @xmath18here @xmath19 and we introduce a small dimensionless parameter @xmath20 which we will use throughout the paper .
notice that at a given energy the velocity of electrons of both chiralities is the same : @xmath21 .
let us discuss the kinematical aspects of the electron scattering at the lateral interface .
an incident ( nonpolarized ) beam comes at angle @xmath22 from the region without a spin - orbit term ( or with a suppressed spin - orbit term ) , denoted as the n region , and when transmitted into the so region splits into two beams of different chirality that propagate at different angles @xmath23 .
figure [ fig : refraction ] illustrates the scattering for the simplest case when @xmath24 .
the conservation of the projection of the momentum on the interface together with eq .
( eq : ppm ) determine the angles of the transmitted and reflected beams ( snell s law ) : @xmath25 where @xmath26 is a momentum of an electron in the @xmath27 region and @xmath28 are the momenta in the so region after passing through the interface .
splits after the refraction into two beams with @xmath13 and @xmath29 chiralities that propagate at angles @xmath30 ( denoted by red and blue colors , respectively ) .
( a ) @xmath31 determines the limited aperture for @xmath14 chirality ; in the angle interval @xmath32 only electrons of @xmath33 chirality can penetrate .
( b ) @xmath34 is an angle for total internal reflection for electrons of the @xmath33 chirality.,scaledwidth=50.0% ] from eqs .
( [ eq : ppm ] ) and ( [ eq : snell ] ) it follows that the so region is more `` optically '' dense for the @xmath33 mode ( i.e. , it has a smaller wave vector ) and less dense for the @xmath35 mode .
correspondingly , the @xmath33 mode is refracted to larger angles than the @xmath35 one , and therefore the @xmath33 mode exhibits a total internal reflection with a critical angle @xmath36 . as to the @xmath35 mode it has a limited aperture in the so region for outgoing angles : @xmath37 remarkably , the angle interval where only the @xmath33 mode can penetrate is not narrow as its width has a square root dependence on @xmath38 .
it follows from snell s law that @xmath39 .
actually , one can reduce @xmath40 even further . with the gates acting selectively on the different regions of the electron gas , @xmath41
, one can alter the position of the bands relative to the fermi level in the n and so regions .
a simple analysis shows that with an increase of @xmath42 ( i.e. , lowering @xmath43 in the normal region ) the angle interval @xmath44 grows and reaches @xmath45 .
starting from this moment the angle interval suitable for spin filtration narrows and eventually becomes @xmath46 , instead of @xmath47 the problem of scattering of electrons at a lateral interface between the two regions with different magnitudes of the bychkov - rashba term has been considered in ref . for the two limiting cases of a sharp , @xmath48 and smooth , @xmath49 interface , where @xmath50 is an electron wavelength and @xmath51 is an effective width of the interface . here
a qualitative description of the analysis of the spin - dependent scattering at the interface will be presented only .
a scattering state of an electron coming from the n region in the state @xmath52 is given by @xmath53 where @xmath54 are spinors corresponding to the @xmath55 chiral modes in the @xmath27 and @xmath56 regions , and @xmath57 and @xmath58 are the amplitudes of the reflected and the transmitted waves . in eq .
( eq : scattering_state ) the interface is at @xmath59 and for simplicity we limit ourselves to the case of the interface between the n and so regions .
a similar expression holds also for @xmath60 which evolves from the incident state @xmath61 the flux of particles impinging on the interface at a given angle @xmath62 can be defined as @xmath63 , where @xmath64 is the cross - sectional area ( width ) of a beam .
the intensity @xmath65of the transmitted flux of a given chirality can be found from the relation @xmath66 here the cosine factors take into consideration the change of the cross - sectional width of the beam as a result of the scattering , while @xmath67 ( we use the fact that for the two chiralities the velocities are the same ) . finally , the intensity of the refracted flux relative to the incident one is equal to @xmath68 for a sharp interface the amplitudes @xmath69 can be found from the continuity conditions , while for a smooth interface one can conduct the analysis of the refraction using a small parameter @xmath70 . in the latter case the electron spin adjusts itself adiabatically to the momentum keeping its polarization in the direction perpendicular to the momentum . and @xmath71 per unit angle as a function of an angle of incidence for sharp ( solid line ) and smooth ( dashed line ) interfaces ( ref . )
.,scaledwidth=30.0% ] it has been shown in ref . that in the course of refraction at the interface with @xmath72 transitions between states of different chirality are strongly suppressed .
for that reason , the drop of the intensities of the transmitted electrons without change of their chirality @xmath73 and @xmath74 , see fig .
[ fig : intensity ] , occurs practically only due to the reflection which becomes decisive only for @xmath75 . in particular
, for a smooth interface the probability of the reflection outside the region of the total internal reflection is almost entirely suppressed because the matrix elements describing reflection are integrals of the rapidly oscillating functions .
consequently , the transmission amplitudes presented in fig .
fig : intensity have almost rectangular shape for a smooth interface . summarizing the results obtained in ref . for
the cases of the sharp and smooth interface one can state that for both the discussed cases an electron in a state of a definite chirality propagates along the classically allowed trajectory for this chirality , while a change of the chirality is very ineffective .
the case of the n - so interface described so far was taken mostly for illustration .
actually , any interface with a change of @xmath6 results in splitting of the trajectories that can be used for the purposes of spin polarization and filtration . in the analysis above , one should replace @xmath76 by the difference of the strength of the spin - orbit interaction @xmath77 across the interface .
in this section we discuss the kinetics of the two - dimensional ( 2d ) electron gas near the boundary between the diffusive and ballistic regions .
we show that within the linear response approximation the equation determining the current distribution function and the equation determining the density and potential profiles are decoupled , except for boundary conditions .
this analysis justifies the concept introduced by landauer:@xcite to determine the transport properties it is enough to consider noninteracting quasiparticles that propagate without interaction , i.e. , ignoring any effects related to the redistribution of the potential due to the coulomb interaction of electrons .
the problem requires , however , fixing up the boundary conditions , which we perform below for a particular geometry .
the kinetic equation describing a stationary flow of electrons by a distribution function @xmath78 is @xmath79 the electron flow is forced by the electric field , @xmath80 .
correspondingly , eq . ( [ eq : kinetic:1 ] ) should be supplied by the poisson equation which when limited to the 2d plane is @xmath81 here the laplacian acts in 3d space , while @xmath82 is the deviation of density of the 2d electron gas from the equilibrium value .
therefore eq .
( [ eq : poisson ] ) should be supplemented with the 3d conditions on @xmath83 that will take care of the 3d environment of the 2d electron gas .
this introduces an element of nonuniversality into this problem .
fortunately , in the linear response approximation the situation is much more tractable and the problem of the current flow of the 2d gas becomes self - contained ( see , e.g. , ref . ) . in the linear response approximation one keeps terms linear in @xmath84 and @xmath85 only : @xmath86 where @xmath87 and @xmath88 is the fermi - dirac equilibrium distribution .
( we assume throughout the paper that the spatial variation of the fermi - energy level as well as other parameters of the electron liquid such as the density of states and the screening length occurs on distances greatly exceeding the wavelength and the mean free path .
for that reason , the spatial variation of the parameters characterizing the electron liquid is ignored below . ) in the presence of the electron - electron interaction there is an additional force that originates from the interaction of the quasiparticles .
this effect has been accounted for by a substitution of the local equilibrium distribution @xmath89 in place of @xmath85 in eq .
( [ eq : linearized ] ) , see ref . .
following refs . and , one can shift the distribution function by a local value of the potential @xmath90 introducing a displacement function @xmath91 @xmath92 , \label{eq : displacement}\ ] ] where the direction of the momentum @xmath93 is given by the angle @xmath94 . in terms of the displacement function
@xmath91 the system of eqs .
( [ eq : linearized ] ) and ( [ eq : poisson ] ) acquires the form @xmath95 where @xmath96 is the inverse screening length of the 2d electron gas , which in the thomas - fermi approximation is equal to @xmath97 ; @xmath98 is @xmath99 averaged over directions of momentum .
thus , although eq .
( [ eq : linearized ] ) together with the poisson equation ( eq : poisson ) constitutes a system of two coupled equations , the potential @xmath100 drops out from eq .
( [ eq:2d - f ] ) governing the function @xmath91 .
given boundary conditions , the displacement function @xmath91 determines the current distribution without any feedback from eq .
( [ eq:2d - phi ] ) : @xmath101 where @xmath102 is the density of states per one spin species , @xmath103 when one is interested in the connection of the current density with the distribution of the potential and density in the 2d electron gas , eq .
( eq:2d - phi ) should to be involved along with the relation connecting @xmath104 with @xmath105 and @xmath106 : @xmath107 .
\label{eq : density}\end{aligned}\ ] ] for good enough conductors the poisson equation reduces to the condition of the electroneutrality@xcite ( which is valid in any dimension ) . under these circumstances
typical distances on which the potential @xmath83 in the left - hand side of eq .
( [ eq:2d - phi ] ) changes are much longer than the screening length @xmath108 correspondingly , @xmath109 and the poisson equation reduces to the condition of the electroneutrality : @xmath110 now we turn to the discussion of the boundary conditions .
let us assume that the diffusive region has a stripe geometry with the @xmath7 axis directed along the stripe ( see fig .
[ fig : basin ] ) . in the diffusive region
the collision term is controlled by the elastic relaxation time @xmath111 , and the kinetic equation acquires the form @xmath112 here we assume that the impurity scattering is of short range nature .
( the angular dependence of scattering of electrons by the impurities does not influence the distribution function of electrons deep inside the diffusive region , but is important for the angular profile of the diffuse emission . ) with the elastic mean free path @xmath113 used as a unit of length , the kinetic equation can be rewritten in terms of the dimensionless variables @xmath114 : @xmath115 here the minus sign in the first term containing @xmath116 appears because @xmath62 is chosen as an angle formed by a momentum with the direction @xmath117 ( note that @xmath117 directs inwards the diffusive region ) . in the current carrying state electrons deep inside the diffusive region
are distributed according to the drude form . correspondingly , at a distance ( counted from the interface ) exceeding few @xmath118 eq .
( [ eq : homogeneous ] ) has a solution : @xmath119 here the factors are fixed in such a way that the current @xmath120 defined as @xmath121 is equal to @xmath122 for the distribution function @xmath123 . with the use of eq .
( [ eq : currentn - p ] ) the physical current @xmath124 can be related to the current @xmath122 defined in eq .
( [ eq : jcurrent ] ) as follows : @xmath125 to obtain the relation between the current @xmath126 and the electric field let us apply a gradient to the both sides of eq .
( [ eq : density ] ) with the distribution function @xmath127 used for @xmath128 .
then , together with the fact that the electric field @xmath129 , one gets the relation : @xmath130 where the conductivity @xmath131 while the diffusion coefficient of charge density , @xmath132 corresponds to the einstein relation , @xmath133 @xmath134 . under the condition of the electroneutrality @xmath135 and @xmath136 .
[ actually eq . (
[ eq : einstein ] ) does not require the drude form , @xmath137 of the distribution .
it holds in the diffusion approximation when the distribution function has only two first harmonics @xmath138 and the functions @xmath139 and @xmath140 vary on scales much exceeding the mean free path @xmath141 to proceed further with eq .
( [ eq:2d - f ] ) , one has to specify the distribution of particles incident from the terminals located on the ballistic side of the device under discussion .
generally , this distribution is not universal as it depends on a particular geometry . for definiteness
, we consider the case when the diffusive region that has stripe geometry runs into a ballistic basin , see fig .
[ fig : basin ] ( a stripe is wide enough allowing the transverse quantization of electrons to be ignored ) . following ref .
, we use the method of characteristics to determine a distribution of particles that impinge onto the diffusive stripe from the ballistic region ( corresponding family of angles @xmath62 will be denoted as @xmath142 ) . in the ballistic region the collision term is controlled by the inelastic relaxation time @xmath143 , and the kinetic equation acquires the form @xmath144 the solution of eq .
( [ eq : kinetballist ] ) for the directions @xmath145 is given by @xmath146 , \label{eq : characteristics}\ ] ] where @xmath147 is the trajectory ( the characteristics ) of an electron that starts at the remote past from a terminal on the ballistic side of the device and reaches the point @xmath148 with the momentum directed along @xmath62 at the moment @xmath149 see dashed lines in fig . [
fig : basin ] .
we assume that the current carrying area widens sharply on the ballistic side of the setup . correspondingly , the current density vanishes inside the ballistic region at a typical distance @xmath150 . under the condition @xmath151 , the integral in eq .
( [ eq : characteristics ] ) is accumulated at distances where electrons are at the equilibrium and the potential @xmath105 is equal to its equilibrium value at a terminal deep inside the ballistic region , @xmath152 then , it follows from eq .
( [ eq : characteristics ] ) that for incoming directions @xmath153 . ) as a boundary condition for the kinetic equation .
the diffusive region has a geometry of a stripe that runs into a ballistic basin .
the current density vanishes inside the ballistic region at a typical distance @xmath154 .
trajectories that start from a terminal on the ballistic side of the device and reach the diffusive stripe are shown by dashed lines.,scaledwidth=30.0% ] this provides us with the boundary condition to be imposed at the interface , @xmath155 on the incoming part of the function : @xmath156 the obtained boundary condition is isotropic .
this remarkable feature is a consequence of the choice of the proper geometry . together with the current distribution deep in the diffusive region given by eq .
( [ eq : drude ] ) , the relation ( [ eq : bcinfinity ] ) constitutes the full set of the boundary conditions needed for the solution of eq .
( eq : kineticelastic ) .
two remarks are now in order to complete the discussion of the boundary conditions .
\(i ) notice that @xmath157 the point is that due to the abrupt change in scattering the solution @xmath91 of eq .
( [ eq : kineticelastic ] ) has a singular derivative near the interface . under these circumstances
, one can not neglect the term @xmath158 in eq .
( [ eq:2d - phi ] ) , and therefore the condition of the electroneutrality is violated in the vicinity of the interface in a strip of a width @xmath159 .
the deviation of the density distribution @xmath160 from the equilibrium at the interface leads to the variation of the potential , and hence @xmath161 differs from @xmath162 .
\(ii ) as eq . ( [ eq : kineticelastic ] ) can be satisfied by @xmath163 , any solution of eq .
( [ eq : kineticelastic ] ) can be shifted by a constant with no consequences for the physical quantities [ at any measurement of the current one registers the difference between fluxes of incoming and outgoing particles and the isotropic part of @xmath128 is cancelled out ] .
it follows from eq .
( [ eq : displacement ] ) that the distribution of particles impinging onto the diffusive region at the interface , @xmath164 , is not affected by a global shift of the potential : @xmath165 $ ] .
this is in full accord with the fact that one is free to shift the potential @xmath83 by a constant . in sec .
[ sec : milne ] and [ sec : spindependentemission ] we choose @xmath166 and correspondingly use @xmath167 as the boundary condition for the function @xmath91 at the interface between the diffusive and ballistic regions .
in this section we present a solution of the classical problem of the diffuse emission ( milne problem@xcite ) ( see fig . [
fig : milne ] ) in the form convenient for the subsequent analysis of sec .
[ sec : spindependentemission ] .
it is given for an arbitrary dimension @xmath51 , but we are interested in the particular case of @xmath168 . .
the ballistic region is to the right , @xmath169.,scaledwidth=30.0% ] let us specify the notation for angle @xmath62 used throughout sec .
sec : milne and [ sec : spindependentemission ] : for each of the two regions , diffusive and ballistic , @xmath62 is chosen as an angle formed by the momentum of an electron with a normal to the interface directed inwards to the corresponding region .
correspondingly , for @xmath170 we adopt the convention@xcite that for electrons propagating away from the interface @xmath171 ( denoted as @xmath172 in what follows ) , while @xmath173 ( denoted @xmath174 ) corresponds to electrons propagating towards the interface .
the general solution of eq .
( [ eq : homogeneous ] ) can be written as a sum of the current carrying drude flow [ see eq .
( [ eq : drude ] ) ] and the currentless `` counterflow '' @xmath175 , i.e. , @xmath176 . at a distance about a few
mean free paths @xmath118 the counterflow @xmath175 approaches an isotropic distribution ( generally , a nonzero constant ) .
for currentless counterflow particles injected into the diffusive region eventually return back to the interface .
hence the distribution function of particles emitted by the diffusive region , @xmath177 , should be determined completely by the distribution function of the injected particles , @xmath178 . due to the linear nature of eq .
( [ eq : homogeneous ] ) , the @xmath179 parts of the distribution function are related linearly @xcite by the angular green s function @xmath180 : @xmath181 here @xmath182 stands for the angular integration in @xmath51 dimensions , while the subscript @xmath33 in the integral indicates that the integration is limited to the directions for which @xmath183 . with the use of eq .
( [ eq : counterflow ] ) one can reexpress the density @xmath184 corresponding to @xmath175 through the incoming part of the distribution only : @xmath185 where @xmath186 .
\label{eq : functionh}\ ] ] the green s function @xmath180 satisfies a nonlinear integral equation which can be derived as follows .
differentiate eq .
( eq : counterflow ) with respect to @xmath187 , express the derivatives @xmath188 through the kinetic equation ( eq : homogeneous ) , and use eq .
( [ eq : counterflow ] ) to eliminate @xmath189 in favor of @xmath190 . in result one
gets @xmath191 where @xmath192 is a total solid angle in @xmath51-dimensions .
as this relation holds for an arbitrary incident distribution the equation for the green s function @xmath180 follows : @xmath193 the counterflow does not carry current and therefore it satisfies @xmath194 = 0 .
\label{eq : nocurrent}\ ] ] this implies that green s function @xmath180 should satisfy the condition @xmath195 that can be verified with the use of eq .
( [ eq : solution - for - s ] ) along with the relation @xmath196 @xmath197 as it has been explained in sec .
[ sec : kinetics ] the geometry of the discussed setup is such that the distribution function in the ballistic region does not contain a component describing particles propagating towards the interface , i.e. , @xmath198 therefore , in the solution @xmath199 one has to counterbalance the part @xmath200 by a proper choice of @xmath201 .
then the emission outside the diffusive region consists of the two contributions : @xmath202 from the drude flow , and @xmath203 which originates from the compensating counterflow .
as @xmath204 is known , the latter contribution can be found with the help of eq .
( [ eq : counterflow ] ) . as a result
, the emission in the milne problem is given by the expression @xmath205 , \label{eq : pre - milne - emission}\ ] ] which can be simplified using eqs .
( [ eq : solution - for - s ] ) and ( eq : condition ) together with the relation @xmath206 .
finally , the angular distribution of the particles emitted from the interface into the clean region is given by @xmath207 this result coincides with the one presented in ref . for the case @xmath208 .
, of the diffuse emission ( red line ) as compared to the drude flow distribution ( blue line ) for @xmath209 .
the function @xmath210 at @xmath211.,scaledwidth=30.0% ] in fig .
[ fig : diffuse_emission ] the function @xmath212 and the angular dependence of the drude flow distribution , @xmath213 , are plotted , both normalized to @xmath209 .
as compared to the drude flow , the diffuse emission distribution in the milne problem has a qualitatively different behavior at large angles .
namely , the distribution flattens and a considerable part of the distribution is transferred to the large angles .
in this section we generalize the solution of the milne problem to the case of spin - dependent reflection at the interface .
we are mainly concerned with the influence of the strong reflection at the angles tangential to the interface on the distribution of the emitted electrons .
we show that the effect of flattening of the angular distribution is the robust property of the solution of the milne problem which becomes even stronger in the presence of such a reflection .
we now concentrate on the calculation of the diffuse emission in the presence of the spin - orbit interaction at the ballistic side of the junction .
as compared to the consideration in the previous section the scattering at the interface now depends on the direction of spin of the incoming electron . on the other hand , electrons after a sequence of random scattering in the diffusive region return to the interface preserving their spin ( see fig .
[ fig : milneso ] ) , as the impurity scattering of electrons in the diffusive region is assumed to be spin - independent . )
is suppressed.,scaledwidth=30.0% ] to describe such scattering , one should introduce a spin density matrix @xmath214 . in the diffusive region
each component of the spin density matrix satisfies the kinetic equation ( [ eq : homogeneous ] ) because in the normal region the hamiltonian is spin - independent .
consequently , one can apply the same angular green s function @xmath180 as in eq .
( [ eq : counterflow ] ) .
we have to describe the counterflow at the interface in the presence of the reflection . for clarification purposes ,
let us imagine an auxiliary line , @xmath215 separating the diffusive region from the n - so interface .
( actually , for a real device there can be a physical , not auxiliary , interval between the diffusion region and the beginning of the interface where @xmath6 and @xmath216 start to vary . ) similarly to the discussion of the milne problem it is assumed that the electron distribution is given by a sum of the drude flow and a counterflow .
the incoming part of the counterflow consists of the contributions : one , @xmath217 counterbalances the incoming part of the drude flow , @xmath218 , while the other , @xmath219 , is determined by the flow incident onto the line @xmath220 as it has been explained is sec .
sec : kinetics and has been already used in eq .
( [ eq : pre - milne - emission ] ) we assume that there are no electrons incident onto the n - so interface from the so region . in the absence of electrons incident from the
so side of the interface the flow incident onto the line @xmath221 comes only as a result of the reflection at the n - so interface [ the latter circumstance explains our choice of a subscript for @xmath222 .
therefore the overall distribution of particles coming from the interface @xmath223 . as
a result of scattering inside the diffusive region @xmath224 transforms into a part of the outgoing counterflow @xmath225 together with the drude flow it yields the following contribution to the outgoing distribution @xmath226 , where @xmath227 is the @xmath228 unit matrix .
the overall distribution of particles emitted from the line @xmath221 and incident onto the n - so interface is given by @xmath229 the second term in this equation is generated by the incident part of the distribution , @xmath230 which in turn is determined by the reflection of @xmath231 at the n - so interface .
the relation between the incident and reflected parts of the distribution should be found from the solution of the scattering problem at the normal side of interface @xmath232 where by @xmath233 we denote a @xmath228 block of the scattering matrix corresponding to the reflection at the normal side of the interface . after substituting eq .
( [ eq : bc - refl - spin ] ) in eq .
( [ eq : emission - spin ] ) one obtains a closed equation for @xmath234 : @xmath235 to analyze the effect of the reflection at the interface on the form of the diffuse emission , we study eq .
( [ eq : emission - spin - with - bc ] ) for the case of a smooth interface . in this limit
the kinematical aspect of the scattering at the interface is most pronounced and not masked by unnecessary complications .
namely , we assume that for electrons of the @xmath35 chirality the probability of the transmission is @xmath236 at all angles , while for electrons of the @xmath33 chirality the probability of the transmission is @xmath236 for @xmath237 and @xmath238 otherwise ; see fig .
[ fig : intensity ] and the related discussion of the smooth interface .
we will use a notation @xmath239 for spinors .
the sign of the first argument indicates the sign of the projection of the momentum of a scattering electron on the direction perpendicular to the interface , while @xmath240 is the sign of the momentum component along the interface ; the superscript @xmath55 denotes the chirality .
also in what follows the integration over @xmath62 will be performed as @xmath241 . in this notation
the reflection part of the scattering matrix for the smooth interface is given by @xmath242 for each direction @xmath174 we introduce the orthogonal basis @xmath243 in such a way that @xmath244 is parallel to the incident momentum @xmath93 ; @xmath245 is perpendicular to the electron gas plane , @xmath246 and @xmath247 is directed along the vector of the polarization of the @xmath33 chirality . for the analysis of eq .
( [ eq : emission - spin - with - bc ] ) it will be convenient to use the four - component bloch vector @xmath248 related to the matrix @xmath249 as follows : @xmath250 where @xmath251 is the unit matrix and @xmath252 are the pauli matrices . in terms of the @xmath248 matrix eq .
( eq : emission - spin - with - bc ) can be rewritten as a set of coupled equations @xmath253\ ; , \label{eq : emission - spin - components-0}\end{aligned}\ ] ] @xmath254 .
\label{eq : emission - spin - components - alf}\end{aligned}\ ] ] remarkably , all components @xmath255 are determined by a single combination @xmath256 .
the combinations @xmath257 describe the probability for an incident electron to be in the @xmath55 chiral state .
the representation of @xmath258 as @xmath259 is a consequence of the choice of the axis @xmath260 along the direction of the spin polarization of an electron in the state of the @xmath33 chirality .
the fact that the right - hand sides of eqs .
( eq : emission - spin - components-0 ) and ( [ eq : emission - spin - components - alf ] ) depend solely on @xmath261 has a simple physical reason : electrons of this chirality only are reflected at the interface .
[ in eq .
( [ eq : emission - spin - components - alf ] ) the component @xmath262 is determined by @xmath261 without any feedback , and @xmath263 vanishes identically in the chosen parametrization . ] to proceed further we have to calculate @xmath264 for both @xmath265 . with the use of the relation that determines the direction of the polarization of a spinor , @xmath266 , the discussed expression
is reduced to @xmath267 .
then eqs .
( eq : emission - spin - components-0 ) and ( [ eq : emission - spin - components - alf ] ) , can be rewritten in the form : @xmath268f_{% \scriptstyle\mathrm{n}}^{+}(-\mu ^{\prime},+1 ) \notag \\ + & [ 1-\cos ( -\varphi ^{\prime}\pm \varphi ) ] f_{\scriptstyle\mathrm{n}% } ^{+}(-\mu ^{\prime},-1)\big\ } , \label{eq : emission - i}\end{aligned}\ ] ] where angles @xmath62 and @xmath269 are defined in fig .
fig : angles . for the difference @xmath270
one gets a homogeneous equation ) . for each angle @xmath271 and @xmath272
there are two possible momentum directions corresponding to a different sign of the @xmath273 component.,scaledwidth=30.0% ] @xmath274\delta f_{\scriptstyle\mathrm{n}}^{+}(-\mu ^{\prime } ) .
\notag\end{aligned}\ ] ] one can check that all eigenvalues of the right - hand side of this equation considered as a kernel of the transformation are less than 1 . therefore we conclude that @xmath275 , i.e. , @xmath276 is a symmetrical function with respect to @xmath277 . finally , this yields for eq .
( [ eq : emission - i ] ) @xmath278 following the same route , it can be found from eq .
( eq : emission - spin - components - alf ) the distribution of particles in the @xmath35 chiral state : @xmath279 equations ( [ eq : numerics ] ) and ( [ eq : numericsminus ] ) can be analyzed using the smallness of @xmath280 , but they can be easily solved numerically . in what follows
we will use the results of numerical analysis for the distribution functions @xmath281 .
ultimately , we are interested in the distributions on the so side of the interface . to achieve this goal
, we have to connect the calculated distributions @xmath258 to the distributions on the so side @xmath282 .
it follows straightforwardly from the liouville s theorem that @xmath283 where @xmath284 and @xmath285 are two directions connected by the snell s law , eq .
( [ eq : snell ] ) . in the subsequent discussion
the notations @xmath286 and @xmath287 will be used instead of @xmath288 and @xmath289 notice that we return to the definition of angles @xmath62 and @xmath290 as it is given in fig .
fig : refraction . finally , in fig .
[ fig : transmission ] our result for @xmath291 is presented . of the @xmath13 ( shown in red ) and `` @xmath35''(shown in blue ) polarization components of the transmitted electrons for @xmath292
as compared with the solution without reflection , @xmath293 ( dashed line).,scaledwidth=30.0% ] we observe that the distribution of particles of the @xmath33 chirality emitted from the diffusive region into the clean one is flat even in the presence of the spin - dependent reflection at the interface .
moreover , at large angles the function @xmath294 is noticeably increased . as a result of the reflection at the interface
there is a redistribution of the population of the particles of the @xmath33 and @xmath35 chiralities .
consequently , the two chiral components of the current , @xmath295 change in the vicinity of the interface , and we are faced with the question of the conservation of the total current @xmath296 .
( we assume for a moment the width of the diffusive stripe @xmath297 to be unity , and until the end of this section will not distinguish between the density of current and the total current . ) to check the current conservation we first calculated numerically the currents @xmath298 with the use of the distribution functions @xmath299 we get that on the n side of the interface the current @xmath300 is equal to the total current @xmath126 deep inside the diffusive region .
further on , it can be shown that the two parts of the physical current @xmath298 do not change after passing the interface . to do this notice that in the so region the chiral components of the physical current are @xmath301 [ for the definition of the physical currents see eqs .
( [ eq : jcurrent ] ) and ( [ eq : jjrelation ] ) ] . having in mind
( [ eq : liouville ] ) and the snell s law in its differential form , @xmath302 , the following relations for the @xmath55 chirality current components on the two sides of the interface can be obtained : @xmath303 and @xmath304 for the @xmath33 chirality component it has to be taken into consideration that the current of the particles impinging on the interface in the interval of angles @xmath305 is canceled out by the flow of the reflected particles .
thus we arrive at the conclusion that the two parts of the physical current @xmath306 do not change after passing the interface , i.e. , @xmath307 . together with the fact of the current conservation in the n region
this implies that the total current in the ballistic so region is equal to the current deep inside the diffusive region .
so far we limit ourselves to the diagonal elements of the density matrix .
this was possible because in the n region the density matrix is diagonal in momentum space , while the current operator is diagonal in spin space .
the situation is different in the so region where the current operator acquires the spin structure , see sec .
[ sec : spinorbit ] .
in addition , the density matrix becomes nondiagonal in momentum space as an electron beam splits into two beams as a result of the refraction at the interface with the inhomogeneous spin - orbit interaction .
we will not discuss this problem for the following reason . as it has been explained in sec .
[ sec : kinetics ] our analysis refers to the case when the current is registered far away from the near - field zone of the orifice , i.e. , at a distance @xmath308 much exceeding the width of the diffusive stripe . at such distances the trajectories of electrons of different chiralities diverge from each other and therefore the nondiagonal in momentum space components of the density matrix become nonlocal in space . since the current operator is local this nonlocality of the density matrix can not show up in the current measurements .
we have analyzed the transport near the interface with the inhomogeneous spin - orbit interaction in terms of the spin density matrix .
the present analysis has been performed under the assumption that the so interaction in the diffusive region is absent ( suppressed completely ) .
in fact , a much weaker condition is sufficient .
it is enough that spin relaxation rate in the diffusive region is controlled by the dyakonov - perel mechanism:@xcite @xmath309 , where @xmath310 is the spin splitting of the energy spectrum in the diffusive region , @xmath311 . under this condition
the spin relaxation of the electrons during the propagation in the diffusive region after the reflection from the interface is negligible .
( the other limit @xmath312 will be considered elsewhere . )
we have verified that the specific property of the solution of the milne problem the existence of the flat distribution with a large portion of electrons emitted at small angles to the interface still holds in the presence of the reflection at the interface . moreover , at large angles the distribution function of electrons of the @xmath33 chirality is noticeably increased , see fig .
[ fig : transmission ] . as it has already been pointed out in sec .
sec : spinorbit , there exists an interval of the outgoing angles , @xmath313 , where only the @xmath33 spin chiral component can penetrate .
together , these observations call upon to exploit the diffuse emission for the purposes of spintronics . in fig .
[ fig : diffuse - filter ] a scheme of a device is presented which can operate as a spin filter for a current passing through the diffusive stripe confined between nonconducting areas a and b. two additional barriers ( or gates ) are set at a distance @xmath308 much larger than the width of the diffusive stripe @xmath297 . within this geometry each collector , @xmath314 and @xmath315 gets spin carriers emitted into the corresponding angular interval @xmath316 .
when @xmath317 , particles of the @xmath33 chirality only can get into the collectors . as a result , the currents inside each of the two collectors are spin - polarized , dominantly along the @xmath318 directions . in the setup under discussion the orifice of the stripe acts as a source of a current with an angular distribution @xmath319 .
the number of electrons of a certain chirality flowing in a direction @xmath290 per unit time ( i.e. the angular flux ) is equal to @xmath320 , where the intensity @xmath321 is related to @xmath322 as @xmath323 .
the factor @xmath324 appears because we consider the flux of the particles outgoing from the orifice at an angle @xmath290 .
the angular dependence of @xmath325 is in full accord with the expressions for the density of the currents @xmath326 ; see expressions under the integrals in eqs .
( [ eq : jplus ] ) and ( [ eq : jminus ] ) . assuming the angular distribution of the emitted electrons to be practically flat ,
the spin - polarized current that can be collected by each of the collectors is @xmath327 . as it has been explained in sec .
sec : spinorbit the width of the angular interval @xmath328 can be as much as @xmath45 . at this point the fraction of spin - polarized current reaches its optimal value @xmath329 all collected electrons have the same chirality that results in a very high level of spin polarization of the current in the collectors . a deviation from the perfect level of spin - polarization
is only due to a small spread of the direction of motion of electrons within an angular interval @xmath316 . in ref . a large
spontaneous spin splitting has been detected in a gate controlled electron gas formed at a in@xmath330ga@xmath331 as / in@xmath330al@xmath331as heterojunction .
the reported splitting corresponds to @xmath332 for such values of @xmath333 one may expect a rather large angular interval @xmath334 that can be used for spin filtration ; @xmath335 . under these conditions ,
the device proposed in fig .
fig : diffuse - filter has the following specifications : a fraction up to @xmath336 of the total current is collected in @xmath314 and is ( almost ) fully spin - polarized along the direction @xmath337 , while the other fraction of @xmath336 is collected in @xmath338 and is spin - polarized along the direction @xmath117 .
after filtration the spin - polarized current can be manipulated similarly to the polarized light in optical devices .
in particular , one can link the spin filter to the switch of the spin - polarized current discussed in ref . . | a lateral interface connecting two regions with different strengths of the bychkov - rashba spin - orbit interaction can be used as a spin polarizer of electrons in two dimensional semiconductor heterostructures [ khodas _ et al .
_ , phys .
rev .
lett . * 92 * , 086602 ( 2004 ) ] . in this paper
we consider the case when one of the two regions is ballistic , while the other one is diffusive .
we generalize the technique developed for the solution of the problem of the diffuse emission to the case of the spin - dependent scattering at the interface and determine the distribution of electrons emitted from the diffusive region .
it is shown that the diffuse emission is an effective way to get electrons propagating at small angles to the interface that are most appropriate for the spin filtration and a subsequent spin manipulation .
finally , a scheme is proposed of a spin filter device , see fig .
[ fig : diffuse - filter ] , that creates two almost fully spin - polarized beams of electrons . |
recent experimental results from the k2k collaboration show a significant deficit of muons in the forward scattering events with respect to the simulations , which could limit the accuracy of the predicted neutrino energy spectrum at the far detector @xcite .
this deficit could be due , among other possibilities , to the overestimation of charged current ( cc ) coherent pion production .
a later search @xcite found no evidence of cc coherent pion production in @xmath0c and obtained an upper limit for the fraction of this process to the total cc interaction well below some estimations based on the rein and sehgal model of ref .
preliminary miniboone results for neutral current @xmath1 production show a similar deficit @xcite in the forward direction when compared with different montecarlo ( mc ) models : nuance @xcite , neugen @xcite and neut @xcite .
the discrepancies between these mc simulations indicate the considerable uncertainties in the theoretical description of coherent pion production .
in addition , miniboone has collected a large set of data for @xmath2 production in @xmath0c induced by muon neutrinos of energies around 0.7 gev @xcite .
a fraction of these pions is created coherently , so that a realistic description of the coherent process is required to analyze and understand these data .
cc coherent pion production has been observed experimentally at higher energies and for several nuclei @xcite .
these experiments were studied theoretically using models based on pcac @xcite which described well the results .
work along the same lines has been carried out at low energies in ref .
the only nuclear medium effect considered in these calculations is the distortion of the final pion .
other approaches have tried to incorporate some additional nuclear medium effects that modify the weak pion production . in ref .
@xcite , the authors used the impulse approximation with undistorted pion waves , but already modified the @xmath3 resonance properties , and therefore the production mechanisms , using an effective @xmath3 mass .
the importance of these nuclear effects was demonstrated there by comparing the results with those obtained using a free @xmath3 .
kelkar et al .
@xcite developed a more sophisticated treatment of the @xmath3 in the nuclear medium , and also included the final pion distortion by solving the klein gordon ( kg ) equation with a pion nucleus optical potential .
this model predicted a very low cross section , compatible with the recent results of ref .
nonetheless , there were several approximations in this work .
on the one side a non relativistic reduction of the hadronic current was done , on the other side all the transverse parts of the amplitude were neglected . whereas these approximations are quite reasonable to get an estimate of the cross section for this process , as discussed in ref .
@xcite , a more complete calculation is required now that new data are becoming available .
more recently , singh et al .
@xcite used similar medium effects on the production mechanisms and improved on the description of the elementary @xmath4 process by using a fully relativistic calculation of this process and including all pieces of the amplitude .
however , the pion distortion was implemented in the eikonal approximation , which is not very reliable for the low energy pions that apparently dominate this reaction .
our aim in this paper is to improve the model of kelkar et al .
@xcite in a similar manner , using a more complete and relativistic elementary amplitude but still keeping a more realistic treatment of the pion distortion which is calculated solving the kg equation . in the following ,
we describe the formalism , including @xmath3 production and decay , medium effects on the production mechanism and the optical potential responsible for the distortion of the pion . in section [ sec
: re ] , we present our results , compare with the available experimental data and make predictions for other nuclei and observables .
finally , our summary and conclusions appear in section [ sec : su ] .
in the cc coherent pion production induced by neutrinos ( @xmath5 ) , the nucleus remains in its ground state .
the process consists of a weak pion production followed by the strong distortion of the pion in its way out of the nucleus .
first , we must consider the elementary process of @xmath2 production ( @xmath6 ) .
there are large experimental uncertainties in this cross section at the low energies relevant to this work , which can be clearly appreciated by comparing , for instance , refs .
@xcite and @xcite .
these discrepancies are ultimately translated into different values for the @xmath3 resonance axial form factors that appear in the theoretical models , and are a source of uncertainty for the calculation of the pion production cross sections . in any case , it seems to be clear that for neutrino energies below 2 gev , this process is dominated by a @xmath7 excitation , @xmath8 or @xmath9 followed by its decay @xcite , and therefore it is larger for protons than for neutrons as discussed below .
the situation is different for @xmath1 production or at higher energies . in that case , the @xmath3 mechanism is not enough to provide a good description of data @xcite .
after the introduction of the elementary model for the reaction we discuss how it is modified in the nuclear medium , due to the density dependent changes of the @xmath3 resonance peak s position and width .
finally , we study the distortion of the pion wave function using a pion nucleus optical potential .
the matrix element for the elementary process @xmath10 is written as @xmath11 with the leptonic current @xmath12 and the hadronic current @xmath13 @xmath14 where @xmath15 is the nucleon mass , @xmath16 and @xmath17 are the rarita schwinger and dirac spinors for the @xmath3 and the nucleon of momentum @xmath18 and @xmath19 , @xmath20 is the momentum transfer , @xmath21 and @xmath22 ( @xmath23 ) are the vector and axial vector transition form factors .
the amplitude for the process on protons , @xmath24 , is related to the previous one by an isospin factor @xmath25 the vector form factors can be related to the electromagnetic ones .
the conservation of the vector part of the current implies that @xmath26 .
the assumption of @xmath27 dominance for the @xmath3 electroproduction amplitude gives @xcite @xmath28 and for @xmath29 we take @xcite @xmath30 except for @xmath31 that can be related to @xmath32 using pcac , there are no other constraints for the axial form factors .
we use the following ones , fitted to neutrino scattering data @xcite @xmath33 } { \left ( 1- { { q^2}\over{m_a^2}}\right)^{-2 } } } , \ ] ] and @xmath34 with @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 gev@xmath40 and for the axial mass we take @xmath41 gev @xcite .
this set of form factors produces a good agreement with weak @xmath3 production data induced by neutrinos on nucleons @xcite .
a new analysis of world electron scattering data @xcite allows to go beyond the @xmath27 approximation and update the vector form factors @xcite .
this information calls for a new extraction of the axial form factors .
some steps in this direction have been recently taken in refs .
@xcite , using models for the elementary reaction which include , apart from @xmath3 excitation , some background terms . for the sake of consistency with our description of pion production on
the nucleon explained previously , based on the @xmath3 dominance , we stick to the set of form factors given above . we still need to take into account the @xmath3 decay into a pion and a nucleon .
we use the following lagrangian to describe the @xmath42 transition @xcite @xmath43 where @xmath44 is the isospin 1/2 to 3/2 transition operator and @xmath45 , @xmath46 and @xmath47 are the @xmath3 , pion and nucleon fields respectively .
the isospin operator produces a factor 1 for the decay @xmath48 and a factor @xmath49 for the decay @xmath50 .
this , together with eq .
( 5 ) , implies that the amplitude for cc @xmath2 production on the proton is three times larger than on the neutron .
the coupling constant @xmath51 is such that the experimental @xmath52 width is reproduced . at the high @xmath3
invariant masses reached with the k2k or miniboone neutrino energies , the finite size of the hadrons becomes relevant .
this can be empirically taken into account with a form factor @xmath53 that modifies the @xmath42 coupling . here
, we adopt the following ansatz @xmath54 with @xmath55 gev , used in coupled channel studies of pion and photoproduction of baryonic resonances @xcite . the new hadronic current which should replace @xmath56 in eq .
( [ eq:1 ] ) , already incorporating the decay into a pion and a nucleon , is given ( for the case of @xmath57 ) by @xmath58 where @xmath59 and @xmath60 are the pion and final nucleon momenta so that @xmath61 .
the @xmath3 propagator is given by @xmath62 where @xmath63 .
the energy dependent @xmath3 width is @xmath64 with @xmath65 the pion momentum in the @xmath3 rest frame . finally , the spin @xmath66 projection operator is given by @xmath67 the @xmath3 properties are strongly modified inside the nuclear medium and have been the subject of intensive study , both experimental and theoretical , for many years , see i.e. @xcite and references therein . in this work
we use the results from refs .
@xcite where the @xmath3 selfenergy is calculated in a many body approach as a function of the local baryon density @xmath68 .
this model has been extensively tested in pion induced inclusive processes @xcite , photonuclear reactions @xcite , electron scattering @xcite and even for coherent pion production induced by photons , electrons or nuclei @xcite . in the nuclear medium
the @xmath3 resonance acquires a selfenergy because of several effects such as pauli blocking of the final nucleon and absorption processes : @xmath69 , @xmath70 or @xmath71 .
the real part can be parametrized as @xmath72 where @xmath73 is the normal nuclear density . here , in addition to the attractive proper selfenergy @xmath74 ,
the effective repulsive contribution , that comes from the iterated @xmath3-hole excitation driven by the landau migdal interaction with @xmath75 , has been added @xcite .
the imaginary part is parametrized by the expression @xmath76 where the terms with the coefficients @xmath77 , @xmath78 and @xmath79 correspond to the processes @xmath70 , @xmath69 and @xmath71 respectively .
the values of @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 and @xmath82 can be found in eq .
( 4.5 ) and table 2 of ref .
the parameterizations are given as a function of the kinetic energy in the laboratory system of a pion that would excite a @xmath3 with the corresponding invariant mass , and are valid in the range 85 mev @xmath83 315 mev .
below 85 mev the contributions from @xmath77 and @xmath79 are rather small and are taken from @xcite , where the model was extended to low energies .
the term with @xmath78 shows a very mild energy dependence and we still use the parameterization from ref . @xcite even at low energies . for @xmath84 above 315
mev we have kept these selfenergy terms constant and equal to their values at the bound .
the uncertainties in these pieces are not very relevant there because the @xmath85 decay becomes very large and absolutely dominant .
finally , the pauli blocking of the @xmath86 decay reduces the @xmath87 free width which now reads as @xmath88 the angular integrals @xmath89 and @xmath90 can be found in appendix b of ref .
this selfenergy is taken into account by making the substitutions @xmath91 and @xmath92 in the propagator of eq .
( [ eq : delprop ] ) .
the hadronic current is further modified in the nucleus , where the nucleons are bound and thus have a momentum distribution . in the impulse approximation , and after summing over all nucleons , we can write it as @xmath93 \sqrt{3}\frac{f^*}{m_\pi } f(p ' ) \tilde{d}(p ' ) p_\pi^\alpha \mathrm{tr}\left\{\bar{u}(0 ) \lambda_{\alpha \beta}{\cal a}^{\beta\mu } u(0)\right\}\,,\ ] ] where @xmath94 is the in - medium @xmath3 propagator and the trace corresponds to the sum over the nucleons spin . in the evaluation of the trace we have taken an average nucleon momentum for the spinors neglecting corrections of order @xmath95 . given the large mass of the nucleus
, the pion energy is taken to be equal to @xmath96 and this defines the modulus of the asymptotic pion momentum .
on the other hand , a momentum of @xmath97 is transferred to the nucleus . in the evaluation of the amplitude
we assume that this momentum is equally shared by the initial and the final nucleon , so that they have @xmath98 and @xmath99 respectively .
this prescription has also been used in refs .
@xcite for coherent @xmath1 photo- and electroproduction .
the approximation is based on the fact that , for gaussian nuclear wave functions , it leads to an exact treatment of the terms linear in momentum of the elementary amplitude and allows for a consistent description of the pion - nucleon and pion - nucleus kinematics @xcite .
pion distortion is taken into account by the replacements @xmath100 and @xmath101 where @xmath102 is an outgoing solution of the kg equation for a pion of asymptotic momentum @xmath103 calculated with the optical potential described below . to test the validity of the eikonal approximation we have also calculated the cross section with the pion wave function @xmath104 where @xmath105 and @xmath106 is the pion selfenergy described in the next section .
the cross section for the coherent process @xmath5 is then given by @xmath107 the pion wave function is the solution of the kg equation with an optical potential .
since most of the produced pions lie in the energy region around the @xmath3 excitation the @xmath3-hole model can be used . in this model , which we take from ref .
@xcite , the optical potential is given by @xmath108 where @xmath109 is the pion energy in the laboratory system and the selfenergy @xmath106 reads as @xmath110 here , @xmath111 is the mandelstam variable of the @xmath86 system and @xmath112 is the pion laboratory momentum .
finally , @xmath113 is given by @xmath114 where the two terms correspond to direct and crossed @xmath3-hole excitations .
the neutron densities @xmath115 are taken from ref .
@xcite and the proton densities @xmath116 are obtained from the parameterizations compiled in ref .
@xcite , both deconvoluted to take into account the finite size of the nucleons as done in @xcite .
the real and imaginary part of the @xmath3 selfenergy @xmath117 and the pauli corrected width , @xmath118 , have been described in section [ sect:22 ] . in coordinate space and for finite nuclei
, this p - wave potential can be cast as @xmath119 \,,\ ] ] where the first term has the standard kisslinger form and the second one accounts for the angular transformation from center - of - mass to laboratory variables .
the @xmath120 dependence in @xmath121 appears via the local density approximation @xmath122 . with this potential
we solve the kg equation and obtain the pion scattering wave function .
the procedure is described in detail in refs .
@xcite . to asses
the quality of the potential one can compare its results for differential cross sections with pion nucleus elastic scattering data .
we obtain an overall good agreement from light to heavy nuclei at the energies relevant for this work .
we can also refer the reader to the article by garcia recio et al .
@xcite , where a very similar potential was considered , using as here the local density approximation and the same values for the @xmath3 selfenergy .
the main differences with that work are their inclusion of a phenomenological s - wave selfenergy , and our inclusion of the @xmath3 crossed term .
we have checked that both pieces produce only minor effects in the @xmath3 resonance region .
we show in fig . [
fig : fig1 ] the differential cross section @xmath123 for cc coherent pion production in @xmath0c at a neutrino energy @xmath124 gev .
this plot shows the effect of the different ingredients of the calculation .
the modification of the elementary reaction mechanism , through the inclusion of the @xmath3 selfenergy in the propagator , already produces a strong reduction of the cross section .
this reduction of around 35% agrees with the results of ref .
the pion distortion further decreases the cross section and moves the peak to lower energies .
this reflects the presence of a strongly absorptive part in the optical potential around the @xmath3 resonance peak .
the final result shows that the pion spectrum is peaked at much lower energies or , equivalently , that the muon energy distribution is peaked at higher energies than for the impulse approximation .
as expected , the eikonal approximation fails for low and intermediate energies , where a better treatment of the pion wave function is clearly required .
our eikonal result differs from the one obtained in ref .
@xcite because of their use of the asymptotic momentum in the amplitude , whereas we take the gradient of the distorted pion wave function , and our more complete treatment of the optical potential ( see eq .
( [ eq : eq ] ) vs. eq . ( 6 ) of ref
. @xcite ) .
gev is also plotted for reference.,scaledwidth=82.0% ] in fig .
[ fig : fig2 ] , we present the same observable averaged over the k2k @xcite and the miniboone @xcite spectra compared with the results for a fixed neutrino energy .
the consideration of high energy neutrinos widens the pion momentum distribution , due to phase space . even when both spectra are quite different and k2k has a larger average neutrino energy ( 1.3 gev ) than miniboone ( 0.75 gev ) , the peak position stays at the same pion momentum , below the @xmath3 resonance .
furthermore , most of the pions have relatively low energies such that the use of the @xmath3-hole model is appropriate .
although the total cross section and the energy distribution are strongly modified by the nuclear medium , the angular distribution of the muons remains relatively unaffected , as can be seen on the left panel of fig .
[ fig : fig3 ] where we compare the result of our model with the impulse approximation , rescaled to match the full calculation at zero degrees . on the right panel ,
the muon angular distributions averaged over the k2k and the miniboone neutrino spectra are shown together with the one obtained for 1 gev neutrinos .
gev and averages over k2k and miniboone fluxes.,scaledwidth=99.0% ] in both cases , the consideration of higher energy neutrinos leads to a narrower angular distribution . as a consequence ,
the miniboone angular distribution is appreciably more forward peaked than the one at 1 gev even when its average energy is lower .
we have also checked that the lower bound on the muon energy used by k2k in ref .
@xcite ( @xmath125 mev / c ) does not modify appreciably this observable .
we show the total cross section as a function of the neutrino energy and for several nuclei in fig .
[ fig : fig4 ] , with the caveat that the pion production model is less satisfactory at high energies .
one reason is that mechanisms , other than the excitation of the @xmath3 resonance , become relevant .
also the pion distortion , based on the @xmath3-hole model , is not appropriate for the high energy pions that can be produced by neutrinos with @xmath126 gev .
nonetheless , the @xmath86 interaction is much weaker at high energies than at the @xmath3 peak and thus distortion effects should be smaller there . in fig .
[ fig : fig5 ] , we present the dependence of the total cross section on the atomic number .
gev as a function of the effective number of participants p = z+n/3 , as explained in the text.,scaledwidth=82.0% ] in this process , the amplitude is the coherent sum of the contributions of all participant nucleons . taking into account the isospin factors
, this implies that the amplitude is proportional to an effective number of participants defined as p = z+n/3 ; here z and n are the number of protons and neutrons respectively .
this could suggest a quadratic dependence of the cross section on p , so that for heavier nuclei , the process could be comparatively larger with respect to incoherent @xmath127 production or other processes .
however , there are several reasons that quench the p dependence .
first , pion absorption is quite strong and forces the reaction to be peripheral .
notice that the cross sections for the inclusive processes are not affected by this .
second , the nuclear form factor is narrower for heavy nuclei , and reduces more the contribution from high momentum transfers .
our result for the integrated cross section for coherent pion production on @xmath0c averaged over the k2k flux is @xmath128 cm@xmath40 .
this value is above the upper limit of @xmath130 cm@xmath40 obtained using the ratio between coherent and @xmath131 , the total cc cross section , from the k2k collaboration @xcite and the value for @xmath131 of their mc calculation . without the experimental threshold for the muon momentum , @xmath132 mev@xmath133 , we obtain @xmath134 cm@xmath40 .
excitation and coherent @xmath2 production.,scaledwidth=82.0% ] there are several factors that might conspire to produce the disagreement between our calculation and the experimental upper bound .
first of all , due to the considerable uncertainty in the experimental data for pion production cross sections on the nucleon , the axial n-@xmath3 form factors are not sufficiently constrained .
a more complete theoretical description of the elementary amplitude with the inclusion of background terms @xcite and heavier resonances @xcite could help , but in order to put such a model on a firm ground , more precise data are required .
the optical potential employed in our calculation is realistic around the @xmath3 peak ( i.e. for pions with 150 - 450 mev / c momenta ) , where most of the strength of the reaction actually concentrates , but has room for improvement both at lower and higher energies .
it is also important to recall that at forward angles , where the coherent process is sizable , nuclear effects play an important role and this may affect the experimental separation of the coherent events from the incoherent ones which , to a large extent , relies on the theoretical models built in the mc simulations . the situation is illustrated in fig .
[ fig : fig6 ] where we plot the muon angular distributions averaged over the k2k flux ( and with @xmath135 mev / c ) for coherent @xmath2 production , together with the main contributions to the total inclusive cc cross section : quasielastic scattering ( qe ) and incoherent @xmath3 excitation . the calculations of the @xmath3 part is performed with the same elementary amplitudes and in - medium effects as given above for the coherent reaction . for the quasielastic process , we have adopted the model of ref .
@xcite but updating the nucleon form factors according to @xcite .
nuclear effects include fermi motion , pauli blocking with a local fermi gas and the renormalization of the weak transition , which is treated as an rpa resummation of particle - hole and @xmath3-hole states .
these nuclear correlations cause a considerable reduction of strength at low @xmath136 ( forward angles ) , as can be seen in fig .
[ fig : fig6 ] , while they are negligible for @xmath137 . therefore , if a model that lacks these correlations is used to extrapolate the data from the region of @xmath138 to forward angles , one might overestimate the qe part , causing an underestimation of the contribution of other mechanisms , like the coherent pion production , to the cross section .
we have studied cc coherent pion production induced by muon neutrinos @xmath139 .
our model takes into account the modification of the production mechanism due to the renormalization of the @xmath3 properties in the nuclear medium and the distortion of the final pion .
the distorted pion wave function is obtained by solving the klein gordon equation with an optical potential based on the @xmath3-hole model .
both effects produce a large reduction of the cross section with respect to the impulse approximation .
the distortion of the pion shifts the peak of the pion energy distributions towards low energies .
the angular distributions are slightly widened by the nuclear medium effects .
the a dependence of the coherent process has been investigated .
we have found that the integrated cross section grows more slowly than one would naively expect due to the strong pion absorption and the effect of nuclear form factors .
therefore , we do not expect that coherent pion production becomes more relevant with respect to the incoherent processes for heavier nuclei .
we have also studied the cross sections averaged over the k2k and miniboone spectra . in the case of k2k
, we find a cross section 30% larger than the upper limit estimated in the experiment .
one should however remember that such upper limit is based on values for the cc cross sections on nuclei ( quasielastic , pion production , etc . ) which are not well known at the rather low energies discussed here .
further improvements in the theoretical description of this reaction will require more precise data for the neutrino nucleon cross sections at low energies that could constrain the axial n-@xmath3 form factors .
as for miniboone , we give predictions for energy and angular distributions which can be useful to compare with and analyze their data . we thank p. novella and j. j. gomez - cadenas for useful discussions .
this work was partially supported by dgi and feder funds , contract bfm2003 - 00856 , by the eu integrated infrastructure initiative hadron physics project contract rii3-ct-2004 - 506078 and by the research cooperation program of japan society for the promotion of science ( jsps ) and spanish csic .
m. h. ahn _
[ k2k collaboration ] , phys .
* 90 * ( 2003 ) 041801 .
m. hasegawa _ et al . _
[ k2k collaboration ] , phys .
* 95 * ( 2005 ) 252301 .
d. rein and l. m. sehgal , nucl .
b * 223 * , 29 ( 1983 ) .
m. o. wascko [ miniboone collaboration ] , nucl .
proc . suppl .
* 159 * , 50 ( 2006 ) .
h. j. grabosch _ et al .
_ [ skat collaboration ] , z. phys .
c * 31 * ( 1986 ) 203 .
p. marage _ et al .
_ [ bebc wa59 collaboration collaboration ] , z. phys .
c * 31 * ( 1986 ) 191 .
allport _ et al . _
[ bebc wa59 collaboration ] , z. phys .
c * 43 * ( 1989 ) 523 .
m. aderholz _ et al . _
[ e632 collaboration ] , phys .
* 63 * ( 1989 ) 2349 . s. willocq
_ et al . _
[ e632 collaboration ] , phys .
d * 47 * ( 1993 ) 2661 .
p. vilain _ et al .
_ [ charm - ii collaboration ] , phys .
b * 313 * ( 1993 ) 267 .
a. a. belkov and b. z. kopeliovich , sov .
j. nucl .
* 46 * ( 1987 )
499 [ yad . fiz .
* 46 * ( 1987 ) 874 ] .
e. a. paschos , a. kartavtsev and g. j. gounaris , phys .
d * 74 * ( 2006 ) 054007 .
h. c. kim , s. schramm and c. j. horowitz , phys .
c * 53 * ( 1996 ) 3131 .
n. g. kelkar , e. oset and p. fernandez de cordoba , phys .
c * 55 * ( 1997 ) 1964 .
s. k. singh , m. sajjad athar and s. ahmad , phys .
* 96 * ( 2006 ) 241801 .
s. ahmad , m. sajjad athar and s. k. singh , phys .
d * 74 * ( 2006 ) 073008
. g. m. radecky _ et al .
_ , phys .
d * 25 * ( 1982 ) 1161 [ erratum - ibid .
d * 26 * ( 1982 ) 3297 ] .
t. kitagaki _ et al .
_ , phys .
d * 34 * ( 1986 ) 2554
. g. l. fogli and g. nardulli , nucl .
b * 160 * ( 1979 ) 116 .
h. c. kim , s. schramm and c. j. horowitz , phys .
c * 53 * ( 1996 ) 2468 .
s. k. singh , m. j. vicente vacas and e. oset , phys .
b * 416 * ( 1998 ) 23 [ erratum - ibid .
b * 423 * ( 1998 ) 428 ] .
l. alvarez - ruso , s. k. singh and m. j. vicente vacas , phys .
c * 59 * ( 1999 ) 3386 .
t. sato , d. uno and t. s. h. lee , phys .
c * 67 * ( 2003 ) 065201 .
p. a. schreiner and f. von hippel , nucl .
b * 58 * ( 1973 ) 333 .
l. alvarez - ruso , s. k. singh and m. j. vicente vacas , phys .
c * 57 * ( 1998 ) 2693 .
t. kitagaki _ et al .
_ , phys .
d * 42 * ( 1990 ) 1331 .
s. j. barish _ et al .
_ , phys .
d * 19 * ( 1979 ) 2521 .
p. allen _ et al . _
[ aachen - bonn - cern - munich - oxford collaboration ] , nucl .
b * 176 * ( 1980 ) 269 .
j. bell _ et al .
_ , phys .
* 41 * ( 1978 ) 1012 .
p. a. zucker , phys .
d * 4 * ( 1971 ) 3350 .
j. bijtebier , nucl .
b * 21 * ( 1970 ) 158 .
l. tiator , d. drechsel , s. kamalov , m. m. giannini , e. santopinto and a. vassallo , eur .
j. a * 19 * ( 2004 ) 55 .
o. lalakulich , e. a. paschos and g. piranishvili , phys .
d * 74 * ( 2006 ) 014009 .
e. hernandez , j. nieves and m. valverde , arxiv : hep - ph/0701149 .
m. benmerrouche , r. m. davidson and n. c. mukhopadhyay , phys .
c * 39 * ( 1989 ) 2339 .
g. penner and u. mosel , phys .
c * 66 * ( 2002 ) 055211 .
l. s. kisslinger and w. l. wang , phys .
lett . * 30 * ( 1973 ) 1071 .
m. hirata , j. h. koch , e. j. moniz and f. lenz , annals phys .
* 120 * ( 1979 ) 205 .
e. oset and w. weise , nucl .
phys . a * 329 * ( 1979 ) 365 .
y. horikawa , m. thies and f. lenz , nucl .
a * 345 * ( 1980 ) 386 .
e. oset , h. toki and w. weise , phys .
rept . * 83 * , 281 ( 1982 ) .
r. a. freedman , g. a. miller and e. m. henley , nucl .
a * 389 * ( 1982 ) 457 .
m. hirata , f. lenz and m. thies , phys .
c * 28 * ( 1983 ) 785 .
e. oset and l. l. salcedo , nucl .
a * 468 * ( 1987 ) 631 .
r. c. carrasco , j. nieves and e. oset , nucl .
a * 565 * ( 1993 ) 797 .
s. hirenzaki , j. nieves , e. oset and m. j. vicente - vacas , phys .
b * 304 * , 198 ( 1993 ) .
fernandez de cordoba , j. nieves , e. oset and m. j. vicente - vacas , phys . lett .
b * 319 * ( 1993 ) 416 .
fernandez de cordoba , e. oset and m. j. vicente - vacas , nucl .
a * 592 * ( 1995 ) 472 .
d. drechsel , l. tiator , s. s. kamalov and s. n. yang , nucl .
a * 660 * , 423 ( 1999 ) . c. garcia - recio , j. nieves and e. oset , nucl .
a * 547 * ( 1992 ) 473 . c. w. de jager , h. de vries and c. de vries , atom .
data nucl .
data tabl .
* 36 * ( 1987 ) 495 .
j. nieves , e. oset and c. garcia - recio , nucl .
a * 554 * ( 1993 ) 509 . c. garcia - recio , e. oset , l. l. salcedo , d. strottman and m. j. lopez , nucl .
a * 526 * , 685 ( 1991 ) .
m. h. ahn _
[ k2k collaboration ] , phys .
d * 74 * ( 2006 ) 072003 .
j. monroe [ miniboone collaboration ] , nucl .
* 139 * ( 2005 ) 59 .
s. k. singh and e. oset , nucl . phys .
a * 542 * ( 1992 ) 587 .
h. budd , a. bodek and j. arrington , arxiv : hep - ex/0308005 . | we analyze the neutrino induced charged current coherent pion production at the energies of interest for recent experiments like k2k and miniboone .
medium effects in the production mechanism and the distortion of the pion wave function , obtained solving the klein gordon equation with a microscopic optical potential , are included in the calculation .
we find a strong reduction of the cross section due to these effects and also substantial modifications of the energy distributions of the final lepton and pion . |
arrow polynomial formulas of polyak and viro @xcite are a computationally attractive way to represent vassiliev s finite type invariants of knots and links @xcite .
the input to such formula is a _ gauss diagram _
@xmath1 of a ( based ) link @xmath2 , @xmath3 in @xmath4 , obtained from any plane projection of @xmath5 , by drawing @xmath6 disjoint oriented circles with basepoints , and marked ( positive / negative ) arrows connecting distinct points on the circles .
points of the @xmath7th circle correspond to values of a parameter for @xmath3 , c.f .
a positive / negative arrow between two points on the @xmath7th and @xmath8th circle is drawn , if and only if , for the corresponding parameter values the plane diagram of @xmath5 has a positive / negative crossing , the arrow points from the underpass to the overpass .
alternatively , we may replace the components of @xmath1 i.e. the circles , with vertical or horizontal oriented segments , we call _ strings _ , assuming that the beginning and end of each string is identified with the basepoint , figure [ fig : gauss - d ] illustrates this situation . an _ arrow diagram _
@xmath9 is an arbitrary unmarked diagram ( i.e. arrows have no signs attached to them ) analogous to a gauss diagram but not necessarily obtained from a plane link diagram .
the arrow diagrams and gauss diagrams can be paired as follows ; for a gauss diagram @xmath10 of the @xmath6component link @xmath5 , a _ representation _ of an arrow diagram @xmath9 in @xmath11 is an embedding of @xmath9 into @xmath11 mapping components of @xmath9 to components of @xmath11 , preserving the basepoints and arrow orientations .
define the _ sign _ of a representation @xmath12 by will be denoted by the greek letters : @xmath13 , @xmath14 , , and arrows of @xmath11 lowercase letters : @xmath15 , @xmath16 , ] @xmath17 where the @xmath18 is a sign of the arrow @xmath19 in @xmath11 .
then @xmath20 stands for the sum @xmath21 taken over all embeddings @xmath22 of @xmath9 in @xmath11 .
an arrowhead of @xmath23 will be denoted by @xmath24 and the arrowtail by @xmath25 .
we write @xmath26 or @xmath27 if @xmath13 , resp .
@xmath15 , has its head on the @xmath7component of @xmath9 , and tail on the @xmath8component .
a formal sum of arrow diagrams @xmath28 with integer coefficients is known as an _ arrow polynomial _
@xcite , and @xmath29 is defined from and by the linear extension . theorem of goussarov @xcite shows that any finite type invariant @xmath30 of knots can be expressed as @xmath31 for a suitable choice of the arrow polynomial @xmath32 .
arrow polynomials of some low degree invariants have been computed in @xcite .
for instance , the second coefficient of the conway polynomial @xmath33 of a knot @xmath34 , represented by a gauss diagram @xmath35 is given by @xmath36{v2-arrow - small.png}}\parbox{\wd0}{\box0}\endgroup}\ , , g_k\rangle$ ] ( c.f .
apart from low degree examples , the arrow diagram formulae are known for : the coefficients of the conway @xcite and the homfly - pt polynomials , @xcite . in the case of _ string links _
@xcite , kravchenko and polyak @xcite introduced a family of link homotopy invariants , called _ tree invariants _ which are closely related to the classical milnor linking numbers , @xcite .
@xmath37{proof - case - i_a.pdf}}}$ ] @xmath37{proof - case - i_c.pdf}}}$ ] @xmath38{borr - string - link.pdf}}}$ ] recall that homotopy or link homotopy is a weaker form of equivalence than isotopy which , in particular , allows self - crossing of strands . specifically , kravchenko and polyak defined a family of _ ( planar ) tree diagrams _ on @xmath6 components with a distinguished component @xmath8 called a _ trunk _ ,
for @xmath39 , @xmath40 and @xmath41 , denote the set of these diagrams with leaves indexed by @xmath42 and a trunk on @xmath8 by(left ) ) , where the index of the trunk is also in @xmath42 . ]
@xmath43 , referring to section [ s : tree - invariants ] for further details . paraphrasing the main theorem of @xcite we have [ thm : z_i;j - invariants ]
let @xmath44 be an @xmath6component string link and @xmath45 its gauss diagram .
consider the following arrow polynomial @xmath46 then @xmath47 defines a link - homotopy invariant of the string link @xmath48 , where @xmath49 and @xmath50 is given by number of arrows in @xmath9 pointing to the right .
invariants @xmath51 are called the _ tree invariants _ and are finite type string link invariants , @xcite .
as such , they must necessarily be determined by the classical @xmath52invariants of string links , @xcite . in their paper @xcite kravchenko and polyak
show that for a particular order of leaves and the trunk , namely for @xmath53 and @xmath8 such that @xmath54 , we have the identity @xmath55 the main theorem of the current paper is a generalization of theorem [ thm : z_i;j - invariants ] to closed @xmath6component links .
analogous result by polyak and viro , theorem 6 in @xcite ( see also @xcite ) , concerns the case @xmath56 .
suppose @xmath5 is a based @xmath6component link and @xmath1 its gauss diagram , for @xmath57 , @xmath58 and @xmath59 the following quantity defines a homotopy invariant of @xmath5 ; @xmath60 where @xmath61 further , the relation of @xmath62invariants and milnor higher linking numbers i.e. @xmath63invariants , @xcite is obtained in the following .
[ cor : bar - mu - bar - z ] for @xmath53 and @xmath64 such that @xmath65 we have @xmath66 recall that the @xmath63invariants are important invariants of homotopy links .
in particular , they are capable of distinguishing @xmath67component homotopy links as proven in @xcite .
the indeterminacy of @xmath63invariants was studied by several authors most notably by levine , see @xcite and references therein .
it is not known however , if an appropriate form of indeterminacy , would provide a complete set of numerical link homotopy invariants .
more recently kotori @xcite , defines @xmath63invariants for turaev s nanophrases @xcite , via magnus expansion adapted to nanophrases .
in contrast , our definition is a direct generalization of kravchenko polyak tree invariants and geared towards further applications in geometric knot theory , @xcite . for
a related connection between trees and milnor invariants of string links one may refer to the recent work in @xcite , and other relevant approaches , e.g. in @xcite and @xcite .
the techniques presented in this paper are geometric and depend on certain natural tree diagram decompositions , related ideas can be found in the work of " ostlund @xcite . as a practical outcome
we also obtain inductive formulas for tree invariants ( previously computed directly for @xmath56 and @xmath68 in @xcite ) which yields a convenient computational algorithm .
the paper is organized as follows ; in section [ s : milnor ] , we review the original construction of @xmath63invariants for links and string links as presented in @xcite .
construction of tree invariants of string links is presented in section [ s : tree - invariants ] , together with definitions of some natural tree stacking operations .
section [ s : tree - invariants ] also includes the above mentioned algorithm for generating arrow polynomials of tree invariants for an arbitrary @xmath6 .
proof of main theorem and corollary [ cor : bar - mu - bar - z ] is provided in section [ s : proof ] . in the last paragraph of this introduction
we mention that the applications of milnor linking numbers are fairly broad ; including distant areas such as topological fluid dynamics or plasma physics , c.f .
@xcite . in the forthcoming paper : @xcite , we show how the arrow diagrammatic formulation of linking numbers can be applied to address a geometric question of freedman and krushkal @xcite , concerning estimates for thickness of @xmath6component links .
both authors acknowledge the generous support of nsf dms 1043009 and darpa yfa n66001 - 11 - 1 - 4132 during the years 2011 - 2015 .
a version of main theorem , without a precise characterization of indeterminacy , was obtained by the second author in his doctoral thesis , @xcite .
in this section we review the construction of milnor invariants , @xcite .
denote by @xmath2 an @xmath6component ordered , oriented link in @xmath69 .
recall that , any diagram of @xmath5 yields the wirtinger presentation of @xmath70 , where the generators are _ meridians _ , one for each arc in the diagram and the relations are derived from the crossings in the diagram , @xcite .
different meridians of a given component are conjugate to each other in @xmath71 , a choice of basepoint on the @xmath8component @xmath72 indicates a preferred meridian @xmath73 . in @xmath71
we also distinguish the _ parallels _ , i.e. push - offs of @xmath72 denoted by @xmath74 satisfying @xmath75 .
consider @xmath76 , the free group generated by the preferred meridians @xmath77 . in @xcite milnor proved that the universal homomorphism @xmath78 descends to an epimorphism of the lower central series quotients : @xmath79 ( in fact , due to the later result of stallings @xcite , @xmath80 is an isomorphism ) . recall that given a group @xmath11 , the lower central series is given as @xmath81 , @xmath82 $ ] ,
@xmath83$], .
thus , for any @xmath50 there exists @xmath84 representing the parallel @xmath74 modulo the @xmath50th stage of the lower central series of @xmath71 ( i.e. @xmath85 ) .
every element of @xmath86 can be regarded as a unit in non - commuting variables @xmath87 . ] @xmath88 via the _ magnus expansion _ , which is a ring homomorphism @xmath89 embedding @xmath86 into @xmath88 as its group of units .
given the @xmath8th parallel @xmath74 , as above , we have the expansion @xmath90 ; s\geq 1 } \mu_{i_1\,\ldots\ , i_s;j}\ , x_{i_1 } x_{i_2}\ldots x_{i_s},\qquad [ n]=\{1,\ldots , n\}.\ ] ] the coefficients @xmath91 are defined for each ordered sequence of integers @xmath92 @xmath93 .
following @xcite , let @xmath94 and @xmath95 in @xcite , milnor proved that , for @xmath96 , @xmath97 , the residue classes @xmath98 are isotopy invariants of @xmath5 , and if the indices in @xmath99 are all distinct , @xmath98 are link - homotopy invariants .
the residue classes @xmath100 are commonly known as the _ milnor linking numbers _ or _ @xmath63invariants _ , and @xmath101 is called the _ indeterminacy_. one obvious property of @xmath63invariants is equivariance under permutations @xmath102 , i.e. @xmath103 where @xmath104 is the link @xmath5 with permuted components ; @xmath105 .
further relations are proven in @xcite , for example cyclic symmetry : @xmath106 in fact , due to cyclic symmetry , we may consider a smaller set @xmath107 from the basic properties of @xmath108 , one obtains @xmath109 for _ string links _ , the construction of @xmath52invariants is completely analogous . recall that an @xmath6component string link ( see figure [ fig : trees](right ) for an example of a string link diagram ) is a smooth embedding @xmath110 , of @xmath6 copies @xmath111, @xmath112 of the unit interval @xmath42 into the cylinder @xmath113 , such that each @xmath114 is anchored at the bottom and top of the cylinder at fixed points @xmath115 , i.e. for each @xmath116 ; @xmath117 each string link @xmath48 can be closed up into a link @xmath118 by adding unlinked connecting strands outside the cylinder @xmath119 , this _ closure _ operation is denoted by @xmath120 in @xcite . as before , one considers the group @xmath121 , this time there is a canonical choice of _ meridians _
@xmath122 represented by loops in @xmath123 based at a fixed point @xmath124 , with @xmath125 .
one also defines _ canonical parallels _ ;
@xmath126 as loops in @xmath127 based at @xmath128 and closed up by fixed arcs in the boundary of @xmath127 , with @xmath129 .
again each parallel @xmath74 has its expansion , modulo the @xmath50stage of the lower central series of @xmath86 and @xmath71 .
differently from the case of closed links , the coefficients @xmath130 do not require the indeterminacy and yield the isotopy invariants of string links , @xcite and for distinct indices in @xmath131 , they define link - homotopy invariants . given a link @xmath118 , obtained as the closure of a string link @xmath48 , we have the following identity , @xcite : @xmath132 the linking numbers @xmath133 significantly depend on the order of integers in @xmath134 ( e.g. , in general @xmath135 . )
in section [ s : intro ] , we introduced planar tree diagrams and tree invariants , in this section we provide their formal definitions , following the source in @xcite .
@xmath39 , @xmath40 and @xmath136 , a _ tree diagram _
@xmath9 with _ leaves _ on components numbered by @xmath42 and a trunk on the @xmath8component , is an arrow diagram which satisfies the following conditions : 1 .
an arrowtail @xmath25 and an arrowhead @xmath24 of an arrow @xmath23 belong to different strings ; 2 .
there is exactly one arrow with an arrowtail on the @xmath7-th string , if @xmath137 and no such arrow if @xmath138 ; 3 .
all arrows have arrowheads on strings indexed by @xmath42 ; 4 .
all arrowheads proceed the ( unique ) arrowtail for each @xmath137 , as we follow the @xmath7-th string along its ( downward ) orientation .
define the _ degree _ of an arrow diagram @xmath9 to be the total number of arrows in the diagram , and note that it is always equal to @xmath139 .
one may visualize any tree diagram @xmath9 , as the rooted tree graph : @xmath140 obtained from @xmath9 by removing the part of each of the components that lies below the corresponding ( unique ) arrowtail ( by * ( d2 ) * and * ( d4 ) * ) .
then , @xmath140 is graph isomorphic to a rooted tree with @xmath141 leaves and the root on the @xmath8component .
every tree is a planar graph , however we will refer to a tree diagram @xmath9 as _ planar _ , if in its planar realization the order of leaves coincides with the initial ordering @xmath142 of the components .
note that the above axioms imply that every arrow @xmath23 is uniquely determined by its coordinates @xmath143 . in figure
[ fig : trees - planar](right ) , there are two planar and a non - planar tree diagrams , together with the trees obtained from each of them . in the remaining part of this paper , for practical purposes , we will make no distinction between @xmath140 and @xmath9 .
@xmath144{a - to - t_a_.pdf}}}$ ] @xmath145{planar.pdf}}}$ ] as in section [ s : intro ] , @xmath146 stands for the set of all planar tree diagrams with leaves on @xmath42 and the trunk on the @xmath8-th component and @xmath147 . [
rem : index - notation ] differently than in @xcite , we will treat the trunk of a tree diagram as one of its leaves .
it makes only a small notational difference and simplifies further considerations .
for instance @xmath148 , where @xmath149 agrees with kravchenko and polyak convention ( which , whenever used , will skip the comma separators ) .
further , we also abbreviate @xmath150;j),\qquad z_{n;j}:=z_{[n];j}=z_{12\ldots\widehat{j}\ldots n;j}.\ ] ] let @xmath45 be a gauss diagram of a string link @xmath48 , see figure [ fig : trees](right ) for an example .
kravchenko and polyak @xcite , consider the following element of a quotient algebra @xmath151 @xcite of @xmath152 where the product is defined via _ tree grafting _ and relations correspond to loday s axioms of disassociative algebras @xcite , @xmath153,\ ] ] where @xmath154 $ ] denotes an equivalence class in @xmath151 , and @xmath155 is called the _ sign of the diagram @xmath9_. [ thm : z_j ] let @xmath48 be an @xmath6component string link and @xmath45 its gauss diagram . then @xmath156 is a @xmath151valued homotopy invariant of @xmath48 .
since the equivalence class @xmath154 $ ] of a tree @xmath9 , with a trunk on the @xmath8-th component , depends only on the set of its leaves , all @xmath157 represent the same equivalence class and the sum given by @xmath158 is a homotopy invariant as stated by theorem [ thm : z_i;j - invariants ] of section [ s : intro ] .
invariants @xmath159 are finite type thus by ( * ? ? ?
* corollary 6.4 ) they can be expressed as polynomials in milnor @xmath52-invariants of string links defined in section [ s : milnor ] . because the index @xmath160 for @xmath161 is ordered and @xmath162 in @xmath163 has the increasing order , by assumption , in general : @xmath164 .
however , if @xmath165 , i.e. @xmath166 has the same order as @xmath42 , then we have the following result [ thm : tree = mu ] let @xmath48 be an ordered string link on @xmath6 strings and let @xmath167 .
then @xmath168 ( see also corollary [ cor : z - mu - prod ] ) . the proof of theorem [ thm : tree = mu ] follows from the fact that the tree invariants @xmath169 and @xmath170 satisfy the same skein relation ( see @xcite ) , and have the same normalization , i.e @xmath171 for all string links @xmath48 with the @xmath8-th component passing in front of all the others .
further relations for tree invariants were proven in ( * ? ? ?
* proposition 4.2 ) and , for convenience , are stated below ( where @xmath57 , @xmath172 ) .
* for any @xmath173 and @xmath174 , @xmath175 , @xmath176 * let @xmath177 be a string link obtained from @xmath48 by reflecting the ordering , i.e. @xmath178 , where @xmath179 . then
, @xmath180 * given a cyclic permutation @xmath181 , let @xmath182 be a string link with renumbered strings according to @xmath183 .
then @xmath184 [ rem : z_i;j - product ] note that ( * s1 * ) shows that a computation of @xmath51 with an arbitrary trunk @xmath185 , can be reduced to two cases : _ 1 . _
the trunk is on the first component and _ 2 . _
the trunk is on the last component . at the end of this section
let us exhibit some lower degree invariants @xmath163 ( for @xmath186 ) , as in @xcite , and express them in terms of @xmath52invariants .
in the case where @xmath187 , we have the following tree invariants ( using our notational conventions ) : @xmath188{z_2_1.pdf}}\parbox{\wd0}{\box0}\endgroup } , g_\ell\bigr\rangle , \quad z_{1,2;2 } ( \ell)=z_{1;2 } ( \ell ) = - \bigl\langle{\begingroup \setbox0=\hbox{\includegraphics[scale=0.27]{z_1_2.pdf}}\parbox{\wd0}{\box0}\endgroup } , g_\ell\bigr\rangle,\ ] ] which agree with the linking number : @xmath189 for diagrams with two arrows , we have @xmath190{z_23_1.pdf}}\parbox{\wd0}{\box0}\endgroup } , g_\ell\bigr\rangle , \quad z_{13 ; 2 } ( \ell ) = - \bigl\langle{\begingroup \setbox0=\hbox{\includegraphics[scale=0.3]{z_13_2.pdf}}\parbox{\wd0}{\box0}\endgroup } , g_\ell\bigr\rangle,\\ & z _ { 12 ; 3 } ( \ell ) = \bigl\langle{\begingroup \setbox0=\hbox{\includegraphics[scale=0.3]{z_12_3.pdf}}\parbox{\wd0}{\box0}\endgroup } , g_\ell\bigr\rangle .
\end{split}\ ] ] as shown in ( * ? ? ?
* proposition 4.2 ) , we have the following identities for the above invariants @xmath191 the first equation is just . in the second identity
, we applied ( * s2 * ) and , in the third identity ( * s3 * ) and with @xmath192 .
the invariants @xmath193 and @xmath194 are not equal and in general differ by a sign and a sum of products of pairwise linking numbers of @xmath48 .
the computation in can be easily generalized as follows .
[ cor : z - mu - prod ] let @xmath57 , @xmath172 , if @xmath173 then @xmath195 in particular for @xmath196 we obtain @xmath197 and for @xmath198 : @xmath199 @xmath200 .
we end this section with a computational example .
let @xmath48 be the string link shown in figure [ fig : trees](right ) along with its gauss diagram @xmath45 , and let us compute @xmath201 using identities in .
note that @xmath45 contains three subdiagrams of the first type , two of which contribute @xmath202 and one of which contributes @xmath203 .
also note that @xmath45 does not contain any subdiagrams of the other two types , as a result @xmath204 .
we begin by introducing a certain useful operation , we refer to as _ tree stacking _ , this operation is related but different from grafting which defines the product of the dissociative algebra @xmath151 , introduced earlier . apart from the fact , that tree stacking is our primary tool in proving main theorem , it is essential to obtain an inductive description of planar trees , which is presented in the second part of this section .
recall , that @xmath205 is the set of planar tree diagrams satisfying axioms ( * d1*)(*d4 * ) of section [ s : tree - invariants ] .
tree stacking operations are indexed by the leaves and defined as @xmath206 where @xmath207 $ ] indexes a fixed leaf of a tree in @xmath205 and @xmath208 is simply obtained by `` gluing '' the trunk of @xmath209 on the @xmath210th leaf of @xmath211 , as shown in figure [ fig : tree - stacking ] . the index @xmath212 of the trunk of @xmath208 is determined as follows : @xmath213 onto the @xmath214 leaf of @xmath215 yields @xmath216 . ]
operations @xmath217 are clearly well defined , i.e. produce planar trees on @xmath218 leaves with the trunk on @xmath212 .
we also set @xmath219 given @xmath220 , define @xmath221 and the _ total number of leaves _ : @xmath222 it will be useful , to keep track of leaves indexing in @xmath211 and @xmath209 after stacking . for @xmath223 , @xmath224 , and @xmath225 , we define the following multindices ( as on figure [ fig : tree - stacking ] ) @xmath226\,\cup\{k+s-1\}\cup\ , [ k+m,\ldots , m+n],\\ i(q;z ) & = [ k,\ldots , k+m-1 ] , \end{split}\ ] ] where @xmath227 indexes the leaves of @xmath228 in @xmath229 .
suppose a tree diagram @xmath230 admits a decomposition ( where the parenthesis are inserted in an arbitrary way ) @xmath231 let @xmath232 $ ] the index of @xmath8th leaf component of @xmath9 ( @xmath8component in short ) , we say that the @xmath8th leaf _ stems _ from @xmath233th factor in @xmath9 , if and only if , it was added at the moment of stacking the @xmath233th factor in .
denote @xmath234 $ index interval of $ b_j$},\\ i(k , b_i ; a ) & = \text{the index of $ k$th leaf of $ b_j$ in the index interval $ [ n]$ of $ a$}. \end{split}\ ] ] note that @xmath235 is only well defined for @xmath236 , i.e. @xmath8th leaf stems from @xmath233 . from
we also obtain the following sign identity , @xmath237 building blocks of trees in @xmath205 are the unique elements of @xmath238 and @xmath239 shown on figure [ fig : elem - trees ] and called _ elementary trees_. and @xmath240.,title="fig : " ] and @xmath240.,title="fig : " ] whenever it is clear from the context the horizontal arrow of @xmath241 or @xmath242 will also be denoted by @xmath241 or @xmath242 .
the following lemma conveys a basic fact that trees can be recursively constructed by adding leaves .
[ lem : a(n;r)-recursive ] suppose @xmath243 , then there is @xmath244 $ ] such that @xmath245 or @xmath246 we will induct with respect to @xmath247 . for @xmath248 ,
the claim is obvious , since @xmath249 , @xmath250 and the only tree in @xmath238 is given by @xmath251 and in @xmath252 the only tree is @xmath253 . for the inductive step , note that for any tree diagram @xmath254 , in its planar realization as shown on figure [ fig : trees - planar ] , horizontal arrows can be pushed up or down , so that there is at most one arrow at each vertical level of the diagram . ordering the arrows , from the top to bottom ,
we let @xmath255 be a top arrow .
the top arrow has the obvious property that there are no arrowheads above @xmath256 along the @xmath7component , and equally are no arrowheads above @xmath257 along the @xmath8component of @xmath9 ( by ( * d2 * ) there can be no arrowtail above an arrowhead ) .
note that @xmath258 generally depends on the choices of heights , i.e. the vertical ordering of horizontal arrows along components .
because of planarity , we claim that @xmath258 is `` short '' i.e. @xmath259 indeed , if @xmath260 and @xmath261 , then there is a @xmath210component , such that @xmath262 . the vertical edge of @xmath210component , would have to intersect the edge @xmath258 , because the first arrowhead along this component is below @xmath256 .
this contradicts planarity of @xmath9 and proves the first case of , the second case of can be shown analogously .
now , thanks to the property ( * d4 * ) , there is no arrowhead / tail along the @xmath8component , below the tail of @xmath258 . therefore
, removing the vertical edge corresponding to that component together with @xmath258 , yields a tree @xmath263 which is a subgraph of @xmath9 .
from the definition of @xmath217 in we obtain or proving the inductive step . clearly , the expansions : @xmath264 , and @xmath265 , for some @xmath263 and @xmath266 may produce isomorphic diagrams .
lemma [ lem : a(n;r)-recursive ] gives an inductive definition of sets @xmath205 as follows @xmath267 the above formula provides an algorithm for computing @xmath205 recursively , and thus obtaining polynomials @xmath268 given in .
figures [ fig : a(2;1)-to - a(3;1 ) ] through [ fig : z_234;1-c ] demonstrate the algorithm , by computing @xmath269 and @xmath270 and yield the arrow polynomial @xmath271 . figure [ fig : a(2;1)-to - a(3;1 ) ] shows how the set @xmath272 is obtained from expansions of the diagram @xmath273 , where @xmath274 , @xmath275 , @xmath276 .
figure [ fig : a(2;1)-to - a(3;1 ) ] shows the corresponding arrow polynomial @xmath277 , see equation .
z_2_1-to - z_23_1 z_23_1-to - z_234_1-a z_23_1-to - z_234_1-b z_23_1-to - z_234_1-c expanding the elements of @xmath269 and eliminating duplicates yields @xmath270 .
observe that the second to last term in figure [ fig : z_234;1-b ] expansion is isomorphic to the first term of expansion in figure [ fig : z_234;1-c ] , and the last term in figure [ fig : z_234;1-b ] expansion is isomorphic to the first term in figure [ fig : z_234;1-a ] .
removing these duplicates yields the following arrow polynomial formula for the @xmath278 invariant , which agrees with the formula obtained by kravchenko and polyak in @xcite , @xmath279{z_234_1-a.pdf}}\parbox{\wd0}{\box0}\endgroup } \\ & { \begingroup \setbox0=\hbox{\includegraphics[scale=0.3]{z_234_1-b.pdf}}\parbox{\wd0}{\box0}\endgroup } , g\bigr\rangle .
\end{split}\ ] ] the next lemma tells us that , with appropriate choices of indexes , operations @xmath217 obey a certain natural shuffle formula ( see and for the index notation ) .
[ lem : shuffle ] let @xmath280 , @xmath281 , @xmath282 . for @xmath283 $ ] and @xmath284 $ ]
, we have @xmath285 where @xmath286 , and the indexes @xmath7 , @xmath8 , @xmath287 , @xmath288 and @xmath289 , @xmath290 are related by @xmath291 respectively for the first and second identity in .
the obvious graph isomorphisms are pictured on figure [ fig : shuffle1 ] and figure [ fig : shuffle2 ] . and
@xmath292 in equation . ] and @xmath293 in equation . ]
with respect to the arrow @xmath13 corresponding to @xmath241 ( left ) or @xmath242 ( right).,title="fig : " ] with respect to the arrow @xmath13 corresponding to @xmath241 ( left ) or @xmath242 ( right).,title="fig : " ] the next lemma shows that any diagram in @xmath294 can be decomposed with respect to a given arrow in @xmath9 as pictured on figure [ fig : arrow - decomp ] which is a natural consequence of lemma [ lem : a(n;r)-recursive ] . [ lem : p , q - decomp ] for any given @xmath230 and @xmath295 , there exist @xmath296 , @xmath297 , and @xmath298 , @xmath299 so that @xmath300 where @xmath301 and @xmath241 ( or @xmath242 respectively ) corresponds to @xmath13 in @xmath9
. the proofs of and are analogous and follow by induction with respect to @xmath6 , thus we provide details only in the case . the base case
@xmath248 follows immediately by choosing @xmath302 . for the inductive step ,
suppose is valid for @xmath6 , and pick any @xmath303 . by lemma [ lem : a(n;r)-recursive ] , one of the following holds for some @xmath244 $ ] : @xmath304 in the case @xmath305 ; if @xmath306 , @xmath26 then @xmath307 , @xmath308 and is obtained by setting @xmath309 , @xmath310 , @xmath311 , @xmath312 .
if @xmath13 is different from @xmath241 in , there must be @xmath313 , which becomes @xmath13 in @xmath9 , after stacking either @xmath241 or @xmath242 , according to .
let @xmath314 be coordinates of @xmath315 , then @xmath287 equals either @xmath316 or @xmath7 . by the inductive hypothesis applied to @xmath263 with @xmath315
we obtain @xmath317 for certain @xmath318 , @xmath319 , @xmath320 , @xmath321 and @xmath322 where @xmath241 in corresponds to @xmath315 . using lemma [ lem : shuffle ] and
, we have the following obvious subcases ; for @xmath323 or @xmath324 as in : * for @xmath325 and @xmath326 we obtain by defining @xmath327 * for @xmath328 we obtain by defining @xmath329 , @xmath330 * for @xmath331 we obtain by defining @xmath332
recall the statement of main theorem ; given a based @xmath6component link @xmath5 in @xmath4 and its gauss diagram @xmath1 , the following quantity defines a homotopy invariant of @xmath5 : @xmath333 for @xmath334 , @xmath335 and @xmath59 , @xmath336 , where @xmath337 apart from the presence of the indeterminacy @xmath338 in the main point of the above formula is that the tree polynomial @xmath163 is evaluated on a gauss diagram of a link @xmath1 rather than a string link @xmath45 as in theorem [ thm : z_i;j - invariants ] .
the proof of main theorem is divided into two parts ; in the first part we analyze the difference @xmath339 , where @xmath340 is a gauss diagram of @xmath5 obtained after moving the basepoint of a component past an over / undercrossing .
a similar basepoint change argument can be found in the work of stlund , @xcite , for @xmath56 , but rather than using stlund s _ diagram fragments _ we perform direct computations with tree diagrams and their suitable tree decompositions as stated in lemma [ lem : p , q - decomp ] . in the second part of the proof ,
we show invariance under the reidemeister moves , where the argument is analogous as the one by kravchenko and polyak in theorem [ thm : z_j ] of @xcite and is included mainly for completeness of the exposition .
let us review our basic notation first ; @xmath10 be a gauss diagram of the @xmath6component based link @xmath5 . for convenience we will always visualize @xmath11 as a string rather than a circular gauss diagram
; see figure [ fig : gauss - d ] .
representation _ of a tree diagram @xmath341 in @xmath11 is an embedding @xmath342 of @xmath9 into @xmath11 mapping components of @xmath9 to the strings of @xmath11 , preserving the endpoints and arrow orientations . recall that arrows of @xmath9 are denoted by the greek letters @xmath13 , @xmath14 , , and arrows of @xmath11 are denoted by lowercase letters : @xmath15 , @xmath16, .
further , the pairing @xmath20 is given by the sum @xmath343 taken over all representations of @xmath9 in @xmath11 .
the following notation for a partial sum will be used frequently @xmath344 where the sum is only over those representations which satisfy a given condition : * cond.*. for instance given an arrow @xmath345 and an arrow @xmath23 , we write @xmath346 note that , for any @xmath13 and @xmath15 , @xmath347 if a gauss diagram @xmath11 has @xmath6components , @xmath348 , @xmath349 , and @xmath350 , @xmath351 , we denote by @xmath352 the subdiagram of @xmath11 obtained by removing components ( together with the adjacent arrows ) of @xmath11 which are not in @xmath42 , and define @xmath353 from the definitions presented in section [ s : intro ] , @xmath354 , equals @xmath355 where @xmath356 is obtained from @xmath9 by simply indexing the leaves and the trunk by @xmath357 .
each step in our analysis is pictured on figures [ fig : basepoint - pass - i ] through [ fig : basepoint - pass - iv ] , where @xmath11 ( left ) and @xmath358 ( right ) differ by a basepoint move , i.e. the arrow @xmath27 in @xmath11 becomes @xmath359 in @xmath358 and the `` dashed '' part of the diagrams is common for both @xmath11 and @xmath358 .
we consider four cases @xmath360@xmath361 as the basepoint of the @xmath7th or @xmath8th string passes through the arrowhead / tail of @xmath15 .
the basepoint passes are denoted by @xmath362 and @xmath363 in each of the figures [ fig : basepoint - pass - i][fig : basepoint - pass - iv ] .
the following result characterizes the difference @xmath364 , for @xmath365 .
[ lem : basepoint - change ] let @xmath11 be a gauss diagram of a based closed link @xmath5 , and consider a gauss diagram @xmath358 obtained from @xmath11 via a moving a basepoint past a crossing in each case : @xmath360@xmath361 . then , we have the following identity @xmath366 for some integer coefficients @xmath367 and where the index sets @xmath368 are given as follows @xmath369\cup [ k , n]\ |\ i=1,\ldots , n-1 ; i+1<k\leq n+1 ; t\in [ 1,i]\text { or } t\in [ k , n]\bigr\}\\ & \qquad \cup \bigl\{[1,k]\cup [ i , n]\ |\ 0\leq k
< i-1;i\leq n ; t\in [ 1,k]\text { or } t\in [ i , n]\bigr\},\\ \mathcal{i}_j & = \bigl\{[k,\ldots , m]\ |\ 1\leq k\leq j;j\leq m\leq n ; k < m\bigr\},\quad j\neq t. \end{split}\ ] ] where @xmath370=\varnothing$ ] for @xmath371 , and @xmath372=\{a\}$ ] . for brevity , let @xmath373 denote @xmath374 .
our basic ingredient in the computation of @xmath375 in is a computation of the difference @xmath376 since the diagrams @xmath11 and @xmath358 differ only by the location of the arrow @xmath15 ( @xmath27 ) there are two basic cases ; first , there exists an arrow @xmath23 , @xmath26 , potentially matching @xmath15 under an embedding of @xmath9 in @xmath11 . in the second case
there is no arrow in @xmath9 with coordinates @xmath143 . in the first case
, we may express , using , as @xmath377 note that the second term of vanishes , i.e. @xmath378 indeed , for any embedding @xmath379 , such that @xmath380 ( i.e. @xmath381 ) .
since @xmath382 and @xmath383 are identical we can define @xmath384 as a composition of @xmath80 and the inclusion @xmath385 .
because @xmath386 ; we conclude @xmath387 which proves .
if a diagram @xmath9 has no arrow @xmath13 matching @xmath15 in @xmath11 , the first term in the sum vanishes , and @xmath388 .
then the same argument as for implies @xmath389 .
therefore , we obtain @xmath390 * case @xmath360 : * suppose @xmath345 , @xmath27 , @xmath260 has its arrowhead closest to the basepoint along the @xmath7th string as pictured on figure [ fig : basepoint - pass - i ] , and @xmath358 is obtained from @xmath11 by applying move @xmath362 ( or equivalently @xmath11 is obtained from @xmath358 via the move @xmath363 ) . and
@xmath391 , where @xmath241 corresponds to @xmath13 , the `` dashed '' vertical arrows , in the left diagram , indicate a possible location for the trunk of @xmath9 indexed by @xmath212.,title="fig : " ] and @xmath391 , where @xmath241 corresponds to @xmath13 , the `` dashed '' vertical arrows , in the left diagram , indicate a possible location for the trunk of @xmath9 indexed by @xmath212.,title="fig : " ] the following cases , define certain disjoint subsets in @xmath294 of tree diagrams which may yield a nonzero right hand side of .
roughly speaking , case @xmath392 concerns the diagrams which have an arrow @xmath26 with @xmath24 at the top of component @xmath7 , such diagrams possibly yield @xmath393 ( figure [ fig : i_b - i_c](left ) ) .
case @xmath394 concerns diagrams having @xmath26 with @xmath24 at the bottom of the @xmath7component , such diagrams possibly contribute to @xmath395 ( figure [ fig : i_b - i_c](right ) ) .
case @xmath396 is a special subcase of @xmath394 and concerns diagrams with a single @xmath26 on the @xmath7component , this can only happen for @xmath397 ( by ( * d1*)(*d4 * ) ) .
* case @xmath396 * , ( diagrams possibly contributing to both @xmath398 and @xmath395 in ): define @xmath399 ( figure [ fig : i_b - i_c](left ) with @xmath312 shows the general form of such diagrams ) . then given a representation @xmath379 , @xmath400
, we may define @xmath384 to differ from @xmath80 only by assigning @xmath359 to @xmath13 , i.e. @xmath401 . because @xmath386 , we obtain @xmath402 , yielding zero on the left hand side of , and as a consequence ; @xmath403 * case @xmath392 * ( diagrams possibly contributing to @xmath398 in ): consider diagrams @xmath404 with an arrow @xmath26 , which admit the following decomposition with respect to @xmath13 ( figure [ fig : i_b - i_c](left ) and equation with @xmath312 , @xmath405 ) @xmath406 where @xmath13 corresponds to @xmath241 .
note , that there is no other arrowhead above the arrowhead of @xmath13 along the @xmath7component of @xmath9 , hence @xmath407 ( see notation in ) and because @xmath408 , @xmath9 can not be in @xmath409 . for a fixed @xmath410
$ ] and @xmath411-i)\cup \{i\}$ ] , denote a set of diagrams @xmath230 admitting the decomposition with @xmath412 and @xmath413 , as @xmath414 . clearly , not every @xmath410 $ ] yields a nonempty @xmath415 ,
we must have @xmath42 and @xmath416 as given by ; with @xmath417 and @xmath407 .
for further convenience , we define @xmath418\ |\ \mathcal{a}_{(\iota.b)}(i , j;i;t)\neq \varnothing\}.\ ] ] using and it follows that @xmath419\cup [ k , n]$ ] , for an appropriate @xmath420 ( see figure [ fig : i_b - i_c ] , the `` shaded '' part of @xmath211 ) and therefore @xmath421\cup [ k , n]\ |\ k > j\}.\ ] ] for @xmath422 , any embedding @xmath379 , @xmath400 has two restrictions @xmath423-i)\cup \{i\}.\ ] ] conversely , consider any embedding @xmath424 , and an arrow @xmath425 . if the image arrow : @xmath426 has its head / tail on the @xmath7th string of @xmath11 it must be below the head of @xmath15 and since @xmath427 , any pair of embeddings @xmath428 , @xmath429 , @xmath430 , yields the `` joint '' embedding @xmath431 since @xmath432 , we have @xmath433 ( using the notation in ) . on the other hand ,
since @xmath13 has a top arrowhead along the @xmath7component of @xmath9 , there must be at least one other arrowhead on the @xmath7component .
axiom ( * d4 * ) implies that there can be no embedding @xmath434 , such that @xmath435 because there is no arrowhead / tail below the head of @xmath359 along the @xmath7th string of @xmath358 ( figure [ fig : basepoint - pass - i](right ) ) .
in turn we obtain @xmath436 , and from , , we may compute , for a fixed @xmath437 , @xmath411-i)\cup\{i\}$ ] : @xmath438 where we used the sign identity , and @xmath141 is determined by with @xmath439 and @xmath212 given in . for any @xmath437 ,
equation implies @xmath440 for this reason we will refer to @xmath211 as a _ free factor _ in the decomposition . setting @xmath441 , we obtain from and : @xmath442 * case @xmath394 * ( diagrams possibly contributing to @xmath395 in ): consider diagrams @xmath230 which have an arrow @xmath13 with @xmath26 , @xmath443 and the arrowhead at the bottom of the @xmath7component of @xmath9 . by lemma [ lem : p , q - decomp ] ,
such diagrams admit the following decomposition with respect to @xmath13 ( figure [ fig : i_b - i_c](right ) and equation with @xmath444 ) @xmath445 where @xmath13 is an arrow in @xmath241 .
note that @xmath13 is a bottom arrow , i.e. there is no other arrowhead below the head of @xmath13 along the trunk of @xmath9 . because @xmath446 , we must have @xmath447 . as in the case
@xmath392 , for a given @xmath448 $ ] and @xmath411-i)\cup \{i\}$ ] , we define the set of tree diagrams in @xmath294 admitting the above decomposition , with @xmath449 , @xmath450 as @xmath451 and @xmath452\ |\ \mathcal{a}_{(\iota.c)}(t , j;i)\neq \varnothing\}$ ] . from it
follows that @xmath453 $ ] , for @xmath454 , yielding @xmath455\ |\ t\leq k < j\}.\ ] ] given @xmath456 , observe that , unless @xmath312 which is covered by case @xmath457 , there is no embedding @xmath379 with @xmath400 , since there is no arrowhead above @xmath15 in @xmath11 along the @xmath212th string of @xmath11 . as a result simplifies to @xmath458 .
using , every embedding @xmath384 , @xmath401 restricts to subdiagrams @xmath459 and @xmath460 , yielding @xmath461 , @xmath462 and @xmath463 , @xmath464 , @xmath465 . because , there is no arrowhead / tail below the head of @xmath359 in @xmath358 , every pair @xmath461 , @xmath463 , @xmath464 , gives an embedding @xmath466 , @xmath401 .
analogous computation as in , shows that @xmath460 is a free factor of decomposition and @xmath467 as in , given @xmath468 we compute @xmath469 where @xmath470 .
coming back to the main argument in case @xmath360 ; note that subsets @xmath471 , @xmath472 and @xmath473 of @xmath294 are disjoint .
if @xmath9 does not belong to their union , then either @xmath305 : it has no arrow @xmath26 , or @xmath474 : there exists @xmath23 such that @xmath26 , but @xmath13 is neither the top nor bottom arrow of @xmath9 .
for such a diagram @xmath9 , we easily obtain from : @xmath389 .
consequently , in case @xmath360 , based on , and we compute @xmath475 proving lemma [ lem : basepoint - change ] in case @xmath360 , where the coefficients @xmath476 ought to be chosen as @xmath477 or @xmath478 above .
* case @xmath479 : * suppose @xmath480 in @xmath11 , @xmath481 has its arrowhead closest to the basepoint along the @xmath7component ( figure [ fig : basepoint - pass - ii ] ) , and @xmath358 is obtained from @xmath11 by applying move @xmath362 ( or equivalently @xmath11 is obtained from @xmath358 via the move @xmath363 ) . the type of diagrams which may contribute to , are `` mirror reflections '' of diagrams described in case @xmath360 .
* case @xmath483 : * diagrams @xmath484 which may yield nonzero both @xmath398 and @xmath395 in need to have an arrow @xmath26 with head on @xmath7component and no other arrow adjacent to that component .
it implies that @xmath7component is the trunk of @xmath9 and @xmath485 .
in fact @xmath9 admits the decomposition on figure [ fig : ii_b - ii_c](right ) with @xmath311 . *
case @xmath486 : * diagrams which may yield @xmath487 in need to have an arrow @xmath26 with a top arrowhead along the @xmath7component and admit the following decomposition with respect to @xmath13 ( figure [ fig : ii_b - ii_c](left ) and equation with @xmath311 ) : @xmath488 where @xmath13 is the arrow in @xmath489 , @xmath301 and @xmath490 . given @xmath410
$ ] , @xmath411-i)\cup \{i\}$ ] , define @xmath491 to be the set of trees in @xmath294 decomposable according to with @xmath492 , and @xmath493 . as usual , we let @xmath494\ |\ \mathcal{a}_{(\iota\iota.b)}(i , j;i;t)\neq \varnothing\}$ ] . from , it follows that @xmath495\cup [ i , n]$ ] , for @xmath496 , giving @xmath497\cup [ i , n]\
|\ k < j\}.\ ] ] since @xmath211 is the free factor in , for any @xmath498 : @xmath499 and since @xmath408 ; we also have @xmath436 yielding @xmath500 * case @xmath482 : * diagrams which may yield @xmath501 in , must have no other head / tail below the head of @xmath13 along the @xmath7component of @xmath9 .
such diagrams admit the following decomposition with respect to @xmath13 ( figure [ fig : ii_b - ii_c](right ) and equation with @xmath444 , @xmath502 ) @xmath503 where @xmath13 is an arrow corresponding to @xmath242 .
let @xmath504 denote the set of trees in @xmath294 decomposable according to with @xmath505 , and @xmath449 , and @xmath506\ |\ \mathcal{a}_{(\iota\iota.c)}(t , j;i)\neq \varnothing\}$ ] . from it follows that @xmath507 $ ] , for @xmath508 ( figure [ fig : ii_b - ii_c](right ) ) and @xmath509\ |\ 1<s+1\leq t\leq n\}.\ ] ] since @xmath510 is the free factor in , for any @xmath511 we have @xmath512 @xmath513 , and since @xmath502 , we also have @xmath514 . using yields @xmath515 based on @xmath483 , and ,
we obtain @xmath516 therefore , lemma [ lem : basepoint - change ] is proven in case @xmath479 , where the coefficients @xmath476 ought to be chosen as @xmath517 or @xmath518 above .
* case @xmath519 : * suppose @xmath480 in @xmath11 , @xmath260 has its arrowtail closest to the basepoint along the @xmath8th string as pictured on figure [ fig : basepoint - pass - iii ] , and @xmath358 is obtained from @xmath11 by applying move @xmath362 ( or equivalently @xmath11 is obtained from @xmath358 via the move @xmath363 ) . as in the previous cases , we need to analyze the right hand side of .
* case @xmath521 : * because there is no arrowhead / tail above the tail of @xmath15 along the @xmath8-component of @xmath11 , any tree @xmath230 for which @xmath393 admits the following decomposition ( see figure [ fig : iii_a - b ] with @xmath311 ) @xmath522 thus there is no other than @xmath13 s arrowhead / tail in @xmath9 along the @xmath8component of @xmath9 , and terms in @xmath398 match those in @xmath395 yielding @xmath523 * case @xmath520 : * if @xmath524 , @xmath9 needs to have an arrow @xmath26 , with the arrow tail on the @xmath8component .
if @xmath13 is the only arrow connected the @xmath8component we obtain .
otherwise , @xmath9 has to decompose as follows ( figure [ fig : iii_a - b ] ) @xmath525 where @xmath13 corresponds to @xmath241 and @xmath526 . given @xmath410
$ ] , such that @xmath185 , @xmath527-i$ ] , define @xmath528 to be the set of trees in @xmath294 which can be decomposed according to with @xmath505 , @xmath449 and let @xmath529\ |\ \mathcal{a}_{(\iota\iota\iota.b)}(i , j;i;t)\neq \varnothing\}$ ] . from , it follows that @xmath530 $ ] , for @xmath531 ; @xmath532 yielding @xmath533\
|\ k\leq j\leq m\leq n ; k < m ; i < k\}.\ ] ] the trunk of @xmath534 , has index : @xmath535 , @xmath536 or @xmath537 .
further , for any @xmath534 we have @xmath538 , and since @xmath539 , we have @xmath514 . therefore @xmath510 is a free factor in and @xmath540 based on @xmath521 and @xmath520 we obtain @xmath541 which ends the proof of lemma [ lem : basepoint - change ] in case @xmath519 .
* case @xmath361 : * suppose @xmath480 in @xmath11 , @xmath481 has its arrowtail closest to the basepoint along the @xmath8th string as pictured on figure [ fig : basepoint - pass - iv ] , and @xmath358 is obtained from @xmath11 applying move @xmath362 ( or equivalently @xmath11 is obtained from @xmath358 via the move @xmath363 ) .
* case @xmath543 : * any tree @xmath9 giving @xmath393 , admits the following decomposition , figure [ fig : iv_a - b ] with @xmath312 ; @xmath544 as in case @xmath521 , terms in the sum @xmath398 match those in @xmath395 yielding .
* case @xmath542 : * if @xmath524 , @xmath9 needs to have an arrow @xmath26 , with the bottom arrow tail on the @xmath8component .
any relevant tree diagram @xmath9 in this case not included in case @xmath543 decomposes as pictured on figure [ fig : iv_a - b ] , i.e. @xmath545 where @xmath13 corresponds to @xmath242 and @xmath526 . given @xmath410
$ ] , such that @xmath185 , @xmath527-i$ ] , define @xmath546 to be the set of trees which can be decomposed according to with @xmath450 , @xmath449 and @xmath547\ |\ \mathcal{a}_{(\iota v.b)}(i , j;i;t)\neq \varnothing\}$ ] . from
, it follows that @xmath548 $ ] , for @xmath549 ; @xmath550 , giving @xmath551\ |\ 1\leq l\leq j\leq k < i , l < k\}.\ ] ] the trunk of @xmath9 , has index @xmath535 , @xmath536 or @xmath537 . for any @xmath552
we obtain @xmath553 , and since @xmath554 , we have @xmath514 . therefore , @xmath460 is a free factor in yielding @xmath555 based on @xmath543 and @xmath542 , we obtain @xmath556 which ends the proof of lemma [ lem : basepoint - change ] in case @xmath361 . figure [ fig : reidemeister - moves ] shows the reidemeister moves : @xmath559 through @xmath560 of link diagrams away from the basepoints and the corresponding local arrow changes in their gauss diagrams .
a goal for this subsection is to prove the following [ lem : reidemeister ] given a gauss diagram @xmath11 of a closed based @xmath6component link , let @xmath358 be a diagram obtained from @xmath11 by applying locally one of the moves @xmath559@xmath560 away from the basepoints of link components . then for any @xmath212 we have @xmath561 * case @xmath559 : * let @xmath11 be a gauss diagram of a based link @xmath5 , and @xmath358 a diagram obtained by the move @xmath559 , i.e. @xmath11 and @xmath358 differ by a single arrow @xmath15 shown on figure [ fig : reidemeister - moves ] . since none of the tree diagrams @xmath562 have arrows with the heads and tails on the same segment , and there exists no embedding @xmath379 , such that @xmath400 for some @xmath23 .
therefore every such embedding is also an into @xmath358 and we obtain @xmath563 , proving @xmath564 for @xmath559 . * case @xmath565 : * the diagram @xmath358 is obtained from @xmath11 by adding locally two parallel arrows @xmath566 and @xmath567 of opposite sign i.e. both arrows have their head on the @xmath7th string and the tail on the @xmath8th string of @xmath358 . by uniqueness of arrows in tree diagrams each embedding @xmath434 can map @xmath26 in @xmath9 to either @xmath566 or @xmath567 , therefore @xmath568 every @xmath384 , @xmath569 , factors through the inclusion @xmath570 , i.e. @xmath571 . since @xmath572 : @xmath379 is an embedding into @xmath11 thus @xmath573 and we obtain @xmath574 .
next consider @xmath384 , @xmath575 , i.e. an embedding which may contribute to @xmath576 , every such embedding can be redefined as @xmath434 , where @xmath577 and @xmath578 , since @xmath566 and @xmath567 have opposite signs in @xmath358 we have @xmath579 .
therefore terms in @xmath576 are in one to one correspondence with terms in @xmath580 , but with opposite sign , yielding @xmath581 . collecting the above facts we obtain : @xmath582 , proving @xmath564 for @xmath565 .
the rearrangement of arrows @xmath586 , @xmath587 and @xmath588 under the @xmath560 move is shown on figure [ fig : reidemeister - moves ] .
given any tree diagram @xmath230 , let @xmath589 note that all three arrows can not be in @xmath9 , as it would contradict planarity of @xmath9 .
let us determine those trees @xmath9
, which can contribute to @xmath590 .
clearly , if @xmath9 contains none of the arrows in any embedding @xmath379 factors through the inclusion @xmath591 and since @xmath592 , @xmath80 also embeds @xmath9 in @xmath358 then , analogously as in the case of * r2 * , we conclude @xmath593 .
if @xmath9 contains exactly one of the arrows in , without loss of generality , suppose @xmath594 , then we have an analog of : @xmath595 the second term of the above sum vanishes by the same reasoning as in the previous paragraph .
the first term also vanishes , because every embedding @xmath379 , @xmath596 can be locally redefined as @xmath384 , @xmath597 , @xmath598 and @xmath386 .
first suppose @xmath9 satisfies @xmath362 , given an embedding @xmath379 , the only possibilities are @xmath305 : @xmath607 , @xmath474 : @xmath400 and @xmath608 or @xmath380 and @xmath609 ; @xmath610 : @xmath400 and @xmath611 .
therefore , we have @xmath612 on the other hand there is no embedding @xmath384 , with @xmath401 and @xmath598 , and becomes @xmath613 it is clear that the first three terms of and agree , giving @xmath614 in order for @xmath9 to contribute to @xmath615 , @xmath9 must contain a subtree @xmath616 shown on figure [ fig : r3-v_a ] , ( because the move is local , and there are no other arrows neighboring @xmath15 , @xmath16 , @xmath210 ) .
denote the set of such diagrams by @xmath617 in case @xmath363 : @xmath618 , fully analogous steps as in
yield @xmath619 for @xmath9 to contribute to @xmath620 , it must contain a subtree @xmath621 shown on figure [ fig : r3-v_a ] , denote the set of such diagrams by @xmath622 in case @xmath623 @xmath606 , there is no embedding @xmath342 , with @xmath596 and @xmath624 , as well as no embedding @xmath625 , with @xmath598 and @xmath626 , yielding @xmath627 as a result we obtain @xmath628 observe that there is a bijection @xmath629 , for a given @xmath630 defined simply by replacing @xmath616 subdiagram with @xmath621 .
for every @xmath630 , and @xmath631 , an embedding @xmath379 may be redefined as @xmath632 , where @xmath633 ( since @xmath634 ) , and @xmath401 and @xmath626 .
since @xmath635 we have @xmath386 ( all arrows @xmath15 , @xmath16 , @xmath210 in @xmath11 , and @xmath359 , @xmath636 , @xmath637 in @xmath358 have the positive sign ) .
in turn we obtain @xmath638 since @xmath639 is a bijection we may conclude that the right hand side of vanishes and @xmath564 proving the claim in case @xmath599 .
* cases @xmath640 : @xmath641 , @xmath642 : @xmath643 , @xmath644 : @xmath645*. for these cases the argument is the same as in case @xmath599 , except @xmath646 and @xmath647 must be used in place of @xmath648 and @xmath649 as shown on figures [ fig : r3-v_b ] , [ fig : r3-v_c ] and [ fig : r3-v_d ] .
* cases @xmath650 : @xmath651 , and @xmath652 : @xmath653 * ; as before the two subcases ( figure [ fig : r3-vi_a ] and figure [ fig : r3-vi_b ] ) are analogous , let us consider @xmath650 in detail , see figure [ fig : r3-vi_a ] . either @xmath601 @xmath602 , @xmath603 @xmath604 or @xmath605 @xmath606 . in case @xmath362 , observe that for any @xmath230 , there is no embedding @xmath379 such that @xmath611 , @xmath654 .
using an analogous identity as equation , we obtain @xmath655 the set of diagrams which may contribute to the left hand side in is denoted by @xmath656 in case @xmath363 there is no embedding which may contribute to @xmath657 thus @xmath658 @xmath659 in case @xmath605 there is no embedding of @xmath9 which may contribute either to @xmath660 or to @xmath661 . as a result
we obtain @xmath662 since diagrams in @xmath663(@xmath664 ) all must contain @xmath665(@xmath666 ) as a local subdiagram , we observe that there is a bijection @xmath667 given by replacing @xmath665 in @xmath668 with @xmath665 . moreover , for any representation @xmath434 there exists a corresponding @xmath632 , defined in the obvious way , and since @xmath669 , we have @xmath579 , and therefore @xmath670 as a result the right hand side of vanishes , and @xmath564 as required . for convenience , and without loss of generality , we will work with @xmath671 $ ] and @xmath672 .
let @xmath5 be an @xmath6component based link and @xmath10 its gauss diagram , recall @xmath673 , according to and .
suppose @xmath674 is a diagram , obtained by moving basepoints along components of @xmath5 in an arbitrary way .
naturally , there is a sequence of diagrams @xmath675 where @xmath676 and @xmath677 , @xmath678 differ just by a single crossing pass as in cases @xmath360@xmath361 of lemma [ lem : basepoint - change ] . from lemma [ lem : basepoint - change ]
we obtain for each @xmath7 : @xmath679 therefore , @xmath680 and substituting in the above equation , we may express @xmath681 as a linear combination of @xmath682 , i.e. @xmath683 where the term @xmath684 is excluded from the sum . applying this step and inductively to terms in the sum above
we conclude that @xmath681 is a linear combination of integers from the set @xmath685 , k\in j\},\ ] ] ( note that the index sets defined in of lemma [ lem : basepoint - change ] , subjected to the above inductive process yield all subsets @xmath686 $ ] with trunks @xmath687 ) . letting @xmath688
, we conclude that @xmath689 and thus @xmath690 is invariant under the basepoint changes .
invariance of @xmath691 under reidemeister moves follows immediately form lemma [ lem : reidemeister ] , if there are no basepoints on the interacting strands locally .
if a move involves basepoints , we may slide them away obtaining a diagram @xmath340 , then using , lemma [ lem : reidemeister ] applied to @xmath340 yields the claim .
the link homotopy invariance of @xmath691 follows from the fact that tree diagram do not have arrows with a head and tail on the same component . thus crossing changes within a given component of @xmath5 ,
do not affect the value of @xmath692 .
recall that we claim @xmath693 after a possible link homotopy of @xmath5 we may assume @xmath694 for some string link @xmath48 , c.f .
@xcite . using and theorem [ thm : tree = mu ] , @xmath695 and from main theorem , @xmath696 therefore , it suffices to prove @xmath697 for that purpose , first observe that by * ( s1 ) * ( equation ) we may consider a subset @xmath698 : @xmath699 , k^+=\max(j ) , k^-=\min(j)\},\ ] ] and @xmath700 , giving us @xmath701 now , corollary [ eq : z - mu - prod ] and cyclic symmetry of @xmath0invariants , yields , proving our claim .
we would like to emphasize that main theorem characterizes the value @xmath702 of the arrow polynomial @xmath163 on an arbitrary gauss diagram @xmath1 of a closed based link @xmath5 , and shows that the residue class is an invariant ( this property is crucial in the forthcoming paper @xcite ) .
the invariant @xmath703 by itself can be defined in an obvious way , via equation and theorem [ thm : tree = mu ] in terms of @xmath704 where @xmath705 is link homotopic to @xmath5 .
d. deturck , h. gluck , r. komendarczyk , p. melvin , c. shonkwiler , and d. s. vela - vick .
generalized gauss maps and integrals for three - component links : toward higher helicities for magnetic fields and fluid flows . , 54(1):013515 , 48 , 2013 .
n. w. evans and m. a. berger .
a hierarchy of linking integrals . in _ topological aspects of the dynamics of fluids and plasmas ( santa barbara , ca , 1991 )
_ , volume 218 of _ nato adv .
_ , pages 237248 .
kluwer acad .
publ . , dordrecht , 1992 . | the paper concerns the _ tree invariants _ of string links , introduced by kravchenko and polyak and closely related to the classical milnor linking numbers also known as @xmath0invariants .
we prove that , analogously as for @xmath0invariants , certain residue classes of tree invariants yield link homotopy invariants of closed links .
the proof is arrow diagramatic and provides a more geometric insight into the indeterminacy through certain tree stacking operations .
further , we show that the indeterminacy of tree invariants is consistent with the original milnor s indeterminacy . for practical purposes
, we also provide a recursive procedure for computing arrow polynomials of tree invariants . |
a major challenge of realistic model building in string theory is the construction of theories with supersymmetry breaking metastable de sitter vacua in their low - energy effective supergravity descriptions .
such models are desirable as they provide a simple explanation for the accelerated expansion of the universe convincingly suggested by cosmological measurements .
unfortunately , de sitter vacua are notoriously difficult to obtain in a string theoretic setting .
it has been made clear by the formulation of several no - go theorems ( see for example @xcite ) why naive attempts of obtaining de sitter models necessarily have to fail in many cases .
a natural possibility is to generate a tiny positive cosmological constant via various types of small corrections to a leading - order minkowski vacuum @xcite .
however , these attempts are often facing difficulties controlling higher - order contributions .
a promising alternative approach is to start with an anti de sitter vacuum and then ` uplift ' this vacuum to a de sitter vacuum by adding additional hard supersymmetry breaking contributions @xcite .
these additional contributions however often complicate the supergravity description of models obtained in this way . despite significant progress in overcoming the difficulties of both the approaches sketched above ( see for example @xcite ) , it still appears useful from a model building as well as from a conceptual point of view to study situations where de sitter vacua arise naturally at the leading order in the low - energy effective supergravity description .
this has for example been done for a broad range of type ii string models in ref .
@xcite . a more general program to identify and study
the main obstacles for the appearance of metastable de sitter vacua has been started in @xcite .
it has been found that the critical parameter is the average mass of the sgoldstinos , which is not allowed to become negative .
this result directly excludes leading order de sitter vacua in situations where only the dilaton or a single khler modulus contributes to supersymmetry breaking , confirming the earlier result in @xcite . the more general case where the moduli space can be factorized into one - dimensional submanifolds and thus the khler potential is given by @xmath1
is also excluded by the same constraint .
this analysis has been extended to two contributing khler moduli , spanning a non - trivial two - dimensional scalar manifold , in the case of compactifications of heterotic string theory as well as orientifold compactifications of type iib string theory in @xcite .
the result of the analysis is an explicit topological constraint on the calabi - yau compactification manifold , which has to be fulfilled for metastable de sitter vacua to exist .
the main goal of this paper is to determine how much of this analysis can be carried over to the more general case where @xmath2 khler moduli contribute to supersymmetry breaking .
we only consider the case of heterotic compactifications .
most of the corresponding results for orientifold compactifications of type iib string theory can be easily obtained by exploiting the duality between these two theories , which in this case simply amounts to a sign change of the relevant quantity @xmath3 defined in eq .
( see for example the appendix of @xcite ) .
this paper is organized as follows . in section
[ sec : metastability ] we briefly recall the constraint on the khler potential in metastable de sitter @xmath0-supergravity coming from the goldstino multiplet . in the low - energy effective supergravity theory obtained from heterotic string theory compactified on a calabi - yau threefold , this constraint can be encoded in the sign of a homogeneous function called @xmath3 .
section [ sect : twoandthree ] studies the properties of @xmath3 and reformulates the task of determining its sign as a tensorial eigenvalue problem .
this formulation is exploited to extend the known result for two - dimensional moduli spaces to the three - dimensional situation .
we then apply the results of section [ sect : twoandthree ] to the study of three - dimensional examples in section [ sec : examples ] .
we conclude in section [ sec : conclusion ] .
we start by briefly reviewing the strategy developed in @xcite to study the existence of metastable vacua with a non - negative cosmological constant in @xmath0-supergravity .
we assume that vector multiplets do not play a significant role in supersymmetry breaking and therefore only consider the chiral multiplets .
the lagrangian of @xmath4 chiral multiplets is completely specified by a single real function @xmath5 of the superfields @xmath6 and their conjugates @xmath7 .
this function can be decomposed as @xmath8 , with the real khler potential @xmath9 and the holomorphic superpotential @xmath10 .
the scalar fields @xmath11 and @xmath12 span a khler manifold with metric @xmath13 the scalar potential is given by @xmath14 where @xmath15 .
the vacuum condition then reads @xmath16 where @xmath17 denotes the khler - covariant derivative .
the corresponding mass squared matrix is given by the vacuum expectation value of the hessian of the scalar potential : @xmath18 here and in the following we omit the @xmath19-brackets for quantities evaluated at the vacuum .
using eq . and
eq . , the entries of @xmath20 can be worked out and are given by ( see @xcite ) @xmath21 where @xmath22 is the riemann tensor of the khler geometry . for the vacuum to be metastable , the mass matrix in eq .
has to be positive definite .
a necessary condition for this to be the case is that the upper - left block @xmath23 is positive definite .
we now consider the projection of this block onto the direction of the goldstino @xmath24 in the space of chiral fermions to isolate the contribution of the goldstino multiplet to the mass matrix : @xmath25 this is a positive combination of eigenvalues of @xmath23 and should therefore be positive if @xmath20 is a positive definite matrix .
using eq . and eq . to calculate @xmath26 more explicitly
, one finds @xcite @xmath27 the superpotential in string compactifications is usually quite complicated due to all sorts of non - perturbative contributions and classical background effects .
thus , we will think of the superpotential as generic , which can in principle be tuned to a suitable value .
the vacuum expectation value of @xmath28 does depend on the superpotential and hence can be varied by varying @xmath10 .
the coefficients in @xmath26 however depend only on the khler geometry and thus the condition @xmath29 for the existence of a supersymmetry breaking metastable vacuum does only depend on the khler potential @xmath9 and gives a constraint on possible khler potentials in viable theories .
the cosmological constant in supergravity theories is given by the vacuum expectation value @xmath30 of the scalar potential .
to incorporate the requirement of a non - negative cosmological constant , it is useful to rewrite @xmath26 as ( see again @xcite ) @xmath31g^ig^{{\bar{j}}}g^mg^{\overline{n}}.\end{aligned}\ ] ] the sign of @xmath32 does only depend on the direction of @xmath28 , not on its length .
assume there is a vector @xmath28 such that @xmath33 . by rescaling @xmath34 one
can always achieve @xmath35 ( i.e. a minkowski vacuum ) at which @xmath36 and thus @xmath37 is still positive . by rescaling @xmath38 a bit further ,
one gets @xmath39 and if the change in @xmath38 is small enough
@xmath40 still holds , proving that the condition for the existence of a metastable de sitter vacuum is satisfied .
conversely , if @xmath41 for all directions of @xmath28 , @xmath26 can never be made positive as long as @xmath42 holds . in summary
: @xmath43 hence , a necessary condition for the existence of metastable de sitter vacua is completely encoded in the sign of @xmath32 . as an example
, we briefly sketch the argument given in ref .
@xcite to exclude classical metastable de sitter vacua for the khler potential in eq . .
the riemann tensor in this case is completely diagonal and its entries are given by @xmath44 where @xmath45 is the scalar curvature of the @xmath46-th one - dimensional submanifold , given by @xmath47 parameterizing the goldstino vector @xmath28 such that @xmath48 , where @xmath49 are real numbers satisfying @xmath50 , we obtain @xmath51 extremizing @xmath32 under the constraint @xmath50 gives for its maximum @xmath52 which vanishes in the no - scale case @xmath53 , thus proving that metastable de sitter vacua do not exist
. we will study compactifications of heterotic string theory on calabi - yau manifolds in the following .
the low - energy limits of these theories are known to give @xmath0-supergravity models @xcite and we can therefore use the constraint described above to restrict their admissible khler potentials .
we will assume that only moduli fields participate significantly in supersymmetry breaking .
this assumption has attracted some phenomenological interest recently , in particular in type iib models @xcite .
nonetheless it should eventually be dropped to make the analysis more universal .
work in that direction has been done in @xcite by the addition of matter fields to the analysis .
it has been found that the study of metastability can ( under some assumptions ) be decomposed into two ` orthogonal ' parts and one of these only involves the moduli fields .
thus , the assumption in this paper is justified in the sense that the results are directly applicable to a subset of the generalized problem . as another simplification
we will also neglect bundle moduli and their interplay with the metric moduli . under this assumption , the moduli space parameterizing the deformations of the calabi - yau manifold @xmath54 consists of the deformations of the complex structure and the deformations of the khler form .
locally , the moduli space then factorizes as @xmath55 where the first factor consists of the khler structure deformations and the second one of the complex structure deformations .
remarkably , it turns out that both @xmath56 and @xmath57 are itself khler manifolds , not only when they are combined to give @xmath58 @xcite .
we will assume that only khler moduli contribute to supersymmetry breaking .
note that , as complex structure moduli and khler moduli are interchanged by mirror - symmetry ( see for example @xcite ) , the analysis for the case where only complex structure moduli take part in supersymmetry breaking would be identical .
the classical volume of the calabi - yau manifold @xmath54 is denoted by @xmath59 .
it holds @xcite @xmath60 where @xmath61 is the khler @xmath62-form .
@xmath61 is a harmonic form and can therefore be written as @xmath63 , where @xmath64 , @xmath65 is a basis of the @xmath66-cohomology group of @xmath54 . in string theory compactifications
an additional geometric structure arises , a real two - form @xmath67 , which is connected to the metric by supersymmetry .
as argued in @xcite , natural local coordinates on the moduli space @xmath56 are given by @xmath68 and the classical volume can be written as @xmath69 the symmetric rank-3 tensor @xmath70 is defined by @xmath71 and consists of the ( real ) calabi - yau triple intersection numbers . in the large - volume limit ,
i.e. if the volume of the calabi - yau is large compared to the string scale , the khler potential of @xmath56 is simply given by @xmath72 note that in particular the dimension of the moduli space @xmath56 ( which we will call @xmath73 throughout this article ) is given by the @xmath62-betti number of the calabi - yau manifold @xmath54 : @xmath74 the computation of the khler metric and the riemann tensor can be found in @xcite and the result reads @xmath75 with @xmath76 note that we dropped the bar above indices referring to derivatives with respect to a complex - conjugated quantity . as the khler potential only depends on real fields , every derivative can be thought of as a derivative w.r.t . a real quantity .
as can be directly checked using e.g. eq .
, eq . and eq . , the supergravity theory satisfies the _ no - scale property _ : @xmath77 in models satisfying the no - scale property it always holds that @xmath78 .
it turns out to be useful to exploit the specialty of the @xmath79-direction by explicitly decomposing the goldstino direction @xmath28 into a part parallel to @xmath79 and a part @xmath80 orthogonal to @xmath79 : @xmath81 using the no - scale property eq .
, the projector onto the orthogonal complement of @xmath79 is given by @xmath82 the function @xmath32 defined in eq .
can be decomposed into a negative - semidefinite part and a part which only involves the orthogonal direction @xmath80 ( in fact this also works in a much broader class of models , see for example @xcite for the general discussion ) : @xmath83 where @xmath84 and @xmath3 are given by @xmath85 the term @xmath86 is always non - positive and thus a necessary condition for the positivity of @xmath32 is the positivity of @xmath3 .
the computation and analysis of @xmath3 is the main goal of this study .
it can be shown ( see @xcite ) that the positivity of @xmath3 and the positivity of @xmath32 are actually equivalent conditions in this case , at least for @xmath87- and @xmath88-dimensional moduli spaces . physically acceptable points @xmath89 on the moduli space
have to satisfy three conditions : @xmath90 where the first condition states the positivity of the volume of the calabi - yau threefold @xmath54 , the second ensures the positivity of the kinetic energy of the moduli fields ( which is also connected to the positivity of the spacetime metric of the compactified dimensions ) and the third condition is the metastability condition .
note that eq . and
eq . are invariant under @xmath91 while eq .
changes its sign .
thus , to determine the physically acceptable regions of the moduli space , it is sufficient to only solve eq . and
eq . and then swap the overall orientation of @xmath79 if necessary to also solve eq . .
note also that one eigenvalue of @xmath92 is always positive due to eq . and
hence @xmath92 is positive definite if and only if @xmath93 for an orthogonal basis @xmath94 , of the orthogonal complement of @xmath79 .
this gives @xmath95 constraints and together with eq .
there is a total of @xmath73 conditions to be satisfied .
we will in the following assume that the points satisfying @xmath96 have already been identified and only study the additional constraint coming from @xmath97 .
to check if @xmath3 can be positive for a suitable goldstino direction @xmath98 , its global maximum as a function of the orthogonal direction @xmath80 has to be determined .
since @xmath3 is a homogeneous function of @xmath80 , we can assume that @xmath80 is normalized : @xmath99 .
to proceed , we fix an arbitrary real orthonormal basis of the subspace orthogonal to @xmath79 , i.e. a set of @xmath95 vectors @xmath100 satisfying @xmath101 in terms of these basis vectors , the projector @xmath102 onto the orthogonal complement of @xmath79 can be written as @xmath103 a general unit vector @xmath80 orthogonal to @xmath79 can be parameterized as @xmath104 with real phases @xmath105 and real @xmath106 satisfying @xmath107 with eq . and eq . @xmath3 in eq .
can be written as @xmath108 , \label{eq : omega_heter_p_gen}\end{aligned}\ ] ] where we defined the symmetric rank 3 tensor @xmath109 and used the abbreviation @xmath110 .
we will in the following assume that @xmath80 is real , i.e. that all complex phases @xmath105 vanish .
this assumption is fully justified for @xmath111 : for @xmath87 , only a global phase is present in @xmath3 , which drops out immediately . for @xmath88 , it can be verified ( see appendix [ sect : complex_phase ] ) that it is safe to set the complex phases @xmath105 to zero in the sense that it does not spoil the validity of the positivity analysis in this case . with vanishing complex phases
, @xmath3 simplifies to @xmath112 here and in the following we use an extension of the abbreviation in eq . :
@xmath113 for vectors @xmath114 , @xmath115 and @xmath116 , while a subscript @xmath117 on the left - hand side stands for a contraction with @xmath100 .
the vector @xmath80 in eq
. can also be parameterized by @xmath118 angles @xmath119 ( e.g. using spherical coordinates ) .
@xmath120 can be supplemented with @xmath118 additional unit vectors @xmath121 such that @xmath122 forms an orthonormal frame orthogonal to @xmath79 .
the projector @xmath102 in eq . can then be expressed as @xmath123 where we defined @xmath124 .
this amounts to the substitution @xmath125 in eq .
, i.e. @xmath126 the derivative of @xmath127 w.r.t .
one of the @xmath118 angles parameterizing @xmath80 ( and thus all @xmath127 ) is again orthogonal to @xmath79 ( because @xmath79 does not depend on any of these angles ) and we can write @xmath128 with some matrix @xmath129 and tensor @xmath130 of coefficient functions whose precise form depend on the parameterization of @xmath80 .
the tensor @xmath130 is antisymmetric in its last two indices ( as can be seen via integration by parts ) : @xmath131 using eq . , eq .
and eq . , a short calculation shows that a critical point of @xmath3 has to satisfy @xmath132 = 0,\quad\alpha=2,\dots,\,p-1.\end{aligned}\ ] ] consequently , a subset of the critical points of @xmath3 is given by the solutions of @xmath133 i.e. by the critical points of @xmath134 ( cf .
we denote these critical points by @xmath135 , @xmath136 and proceed by assuming that the global maximum of @xmath3 is indeed contained in this set of critical points .
this assumption can be explicitly verified in the three - dimensional case ( see sect . [
sec : fourth_crit ] in the appendix ) . at the critical points given by eq .
, @xmath3 reads @xmath137 we will show that solving eq .
is equivalent to solving the tensorial eigenvalue problem @xmath138 where @xmath139 a nonzero vector @xmath115 is called a _ tensorial eigenvector _ of @xmath140 if there exists a @xmath141 ( the corresponding _ tensorial eigenvalue _ ) such that @xmath115 solves eq . .
the role of the determinant in linear ( matrix ) eigenvalue problems is now played by the discriminant ( which is therefore sometimes called the _ _ hyperdeterminant _ _ ) @xmath142 $ ] , defined as the minimal - degree homogeneous polynomial in the tensor components @xmath70 which satisfies @xmath143=0\leftrightarrow \exists v^i\neq0:\ , d_{ijk}v^jv^k=0.\end{aligned}\ ] ] these properties fix the discriminant uniquely up to a normalization .
the normalization is typically chosen such that @xmath144=1,\end{aligned}\ ] ] where @xmath145 is the unit tensor . closed form expressions for the discriminant are known for small @xmath73 ( see e.g. eq . for @xmath87 and ref .
@xcite for @xmath88 ) .
formulas to systematically compute discriminants for larger @xmath73 can be found in ref .
@xcite , but the resulting expressions are extremely lengthy .
the discriminant is an invariant of @xmath70 , meaning that under the transformation @xmath146 with some @xmath147-matrix @xmath148 the discriminant transforms as @xmath149 \to \operatorname{\delta}\left[d_{ijk}'\right ] = \left(\det u\right)^{\frac{3}{p}\deg \operatorname{\delta}\left[d_{ijk}\right ] } \operatorname{\delta}\left[d_{ijk}\right],\end{aligned}\ ] ] where @xmath150=p\cdot 2^{p-1}\end{aligned}\ ] ] denotes the degree of the discriminant in the tensor entries @xmath70 .
equation implies that all tensorial eigenvalues are roots of the _
characteristic polynomial_. @xmath151=0.\end{aligned}\ ] ] together with eq .
this implies that up to normalization the discriminant is equal to the product of all tensorial eigenvalues : @xmath152\cdot\prod_{n=1}^{p\cdot 2^{p-1 } } \lambda_n= \operatorname{\delta}\left[e^kd_{ijk}\right],\end{aligned}\ ] ] where @xmath153 are the solutions of eq . .
we now prove the following statement : if @xmath80 satisfies @xmath154 then three solutions of the eigenvalue problem @xmath155 with the right - hand side defined in eq . ,
are given by @xmath156 conversely , solutions of the eigenvalue problem of the form @xmath157 give critical points of @xmath134 satisfying eq . .
eq . and
eq . are obtained by plugging the ansatz in eq . into the eigenvector equations and multiplying these with @xmath79 and @xmath80 respectively .
to prove that @xmath115 actually solves the eigenvector problem , we introduce a lagrange multiplier @xmath158 to implement the constraint @xmath159 and consider the function @xmath160 extremizing this function is equivalent to extremizing @xmath134 under the constraints in eq . .
differentiating with respect to @xmath80 gives @xmath161 , \label{eq : dnnn_crit_lagrange}\end{aligned}\ ] ] where in the second line we used both constraints from eq . and
@xmath162 , which follows from multiplying the first line with @xmath163 .
multiplication of @xmath164 with an arbitrary vector @xmath165 orthogonal to @xmath79 and use of eq . , and gives @xmath166 which proves the claim .
the converse statement is proven by deducing eq . from eq . .
the two - dimensional case @xmath87 has been studied in ref . @xcite . if @xmath87 , the subspace orthogonal to @xmath79 is 1-dimensional and the ( up to orientation ) only real unit vector orthogonal to @xmath79 is @xmath167 the orthogonal part @xmath80 contains no free parameters and no extremization needs to be performed .
@xmath3 can be obtained directly by a somewhat tedious calculation .
the result is @xmath168}{\det g^3},\end{aligned}\ ] ] where @xmath142 $ ] is a homogeneous polynomial in the intersection numbers defined by the formula @xmath169 : = d_{111}^2d_{222}^2 + 4d_{111}d_{122}^3 + 4d_{222}d_{112}^3 - 3d_{112}^2d_{122}^2 -6d_{111}d_{222}d_{112}d_{122}.\end{aligned}\ ] ] this polynomial is the discriminant of the homogeneous polynomial @xmath170 .
note that the usual definition of the discriminant of a binary cubic includes an additional factor of @xmath171 . in physical regions of the moduli space
, @xmath172 has to hold . in these regions
, @xmath3 is positive if and only if the discriminant @xmath142 $ ] is positive .
the same result can be found by using the tensorial eigenvalue problem in eq . .
note that in the @xmath87-dimensional case , @xmath80 in eq .
is fixed by eq .
and together with @xmath79 it spans the whole vector space .
thus , eq . , eq . and eq . completely specify all tensorial eigenvectors with non - vanishing orthogonal component .
the eigenvalues are given by the solutions of eq .
( where now @xmath173 ) and their product satisfies @xmath174 there is one additional eigenvalue @xmath175 with an eigenvector in @xmath79-direction : @xmath176 where we used eq . and the definition of @xmath177 in eq . . using eq .
, we find @xmath178}{\operatorname{\delta}\left[i_{ijk}\right]}.\end{aligned}\ ] ] using eq . to calculate the discriminant of @xmath177 gives @xmath179 = \frac{4}{3}\det g^3,\end{aligned}\ ] ] which back in eq .
reproduces the result in eq . .
we now consider the three - dimensional situation . to simplify the notation , we explicitly parameterize @xmath80 by @xmath180 with real orthonormal vectors @xmath181 and @xmath182 as in eq . .
the real unit vector @xmath183 orthogonal to @xmath79 and @xmath80 is then given by @xmath184 according to eq .
, a local extremum of @xmath3 has to satisfy @xmath185 i.e. either @xmath186 or @xmath187 holds .
it can be explicitly checked ( the argument can be found in section [ sec : fourth_crit ] in the appendix ) that the second possibility , which is in fact a linear equation in @xmath188 , can be discarded in the positivity analysis , as the corresponding critical point is not the global maximum of @xmath3 .
hence , we are left with the task of determining the solutions of @xmath189 i.e. finding the critical points of @xmath190 .
this equation is cubic in @xmath188 .
its solutions can be determined analytically , though the explicit results unfortunately are not very illuminating .
the tensorial eigenvalue formulation derived in section [ sec : tens_ev_prob ] however turns out to be more suitable for studying the properties of the critical points of @xmath3 .
we now exploit the formulation of the maximization problem for @xmath3 in terms of the tensorial eigenvalue problem eq . to obtain the generalization of eq . in the @xmath88-dimensional case . for this
, we explicitly identify all tensorial eigenvalues specified by eq . .
as in the two - dimensional case , one eigenvalue is always given by @xmath191 with an eigenvector in @xmath79-direction : @xmath192 the other eigenvalues are given by the solutions of @xmath193 where @xmath134 is evaluated at one of its three critical points .
for every critical point of @xmath134 , this equation gives three tensorial eigenvalues .
if the critical point is real , exactly two of them will be complex and the third will be real .
in addition , we have @xmath194 and thus @xmath97 if and only if the real eigenvalue is negative .
the product of the three eigenvalues from the @xmath195-th critical point @xmath196 satisfies @xmath197 the three critical points of @xmath134 give @xmath198 eigenvalues via eq . and together with the one in eq .
we found @xmath199 eigenvalues .
thus , the characteristic polynomial in eq .
is of degree @xmath199 .
however , the discriminant is of degree @xmath200 in the tensor components and we must conclude that the right - hand side @xmath177 is ( doubly ) degenerated : @xmath201=0.\end{aligned}\ ] ] this spoils the validity of eq . , which we otherwise could have used to derive a generalization of eq .
. to repair this flaw , the eigenvalue problem eq
. has to be regularized , i.e. we have to replace the tensor on the right - hand side : @xmath202 such that for @xmath203 @xmath204 \not= 0.\end{aligned}\ ] ] one possibility is @xmath205 this substitution deforms the eigenvalues found above only by terms of order @xmath206 but introduces two additional eigenvalues .
these can be calculated by an expansion in @xmath206 for the eigenvector @xmath115 and the eigenvalue @xmath26 of the form @xmath207 plugging eq . into the eigenvector equations
gives for the product of the two new eigenvalues @xmath208 where @xmath209 putting everything together and using eq .
we finally find @xmath210\frac{|{\mathrm{i}}c_1+c_2|^2}{(d_{111}-3d_{122})^2+(d_{222}-3d_{112})^2}\left[\lim_{\varepsilon\to 0}\frac{\varepsilon^2}{\operatorname{\delta}\left[i_{ijk}^\varepsilon\right]}\right],\end{aligned}\ ] ] where @xmath211 are the three critical points of @xmath134 .
it remains to compute the quantity @xmath212\right]$ ] .
this can be done either via a brute - force approach using the explicit expressions for the discriminant derived in @xcite or via a perturbative expansion for the eigenvectors of @xmath213 . in either way
, one finds for the leading order result in @xmath206 @xmath204 = \varepsilon^2\frac{4 ^ 3}{3 ^ 4}|{\mathrm{i}}c_1+c_2|^2 \det g^6+\mathcal{o}(\varepsilon^3),\end{aligned}\ ] ] which gives @xmath214}{(d_{111}-3d_{122})^2+(d_{222}-3d_{112})^2}.\end{aligned}\ ] ] equation is the three - dimensional generalization of the two - dimensional result eq . .
unfortunately , it has less predictive power as it is only a statement about the _ product _ of a subset of the critical points of @xmath3 , though we know that this subset contains the global maximum . however , one important conclusion can be drawn , namely @xmath215 > 0
\quad\rightarrow\quad \omega = -1+\frac{3}{2}\left(d_{nnn}^2+d_{nnm}^2\right ) > 0\end{aligned}\ ] ] at at least one critical point of @xmath3 , providing that @xmath96 is satisfied .
note however that the converse does not necessarily hold : for negative @xmath142 $ ] either one or all critical points of @xmath3 on the left - hand side of eq . can be negative and the number of negative critical points may even vary on the moduli space .
to verify eq . and to study the possible existence of a converse of eq .
we performed a numerical study .
note that the discriminant @xmath142 $ ] can be written as ( see @xcite ) @xmath149 = t^2\left[d_{ijk}\right ] - s^3\left[d_{ijk}\right],\end{aligned}\ ] ] where @xmath216 $ ] and @xmath217 $ ] denote the _ _ aronhold invariants _ _ is sometimes defined with an additional factor of @xmath218 . ] , well - known invariants of cubic polynomials in three dimensions of degree 4 and 6 in the tensor components respectively ( see @xcite for a modern exposition ) . a numerical code has been used to randomly generate sets of intersection numbers and to classify the resulting models as either _ purely positive _ , meaning that every physical point on the moduli space fulfills @xmath219 or as _ partially positive _ , meaning that only a subset of the physical points allow for a positive @xmath3 . in the generic case , @xmath220 ,
no model has been found which would qualify for a _ purely negative _ classification , i.e. which does not allow metastable de sitter vacua at all . plotting all generated models in the @xmath221-plane ( see fig .
[ fig : heterotic ] ) shows the correctness of eq . and demonstrates that its converse does not hold : all generated models with @xmath142<0 $ ] allow metastable de sitter vacua at a proper subset of the moduli configurations satisfying the basic requirement @xmath96 .
-plane showing the metastability classification of @xmath222 randomly generated heterotic string models .
note that there are no purely negative points . ]
the result in the last section can in principle be generalized to arbitrary @xmath73 ; the explicit calculations however become quite involved and have not yet been carried out completely . in this section
we briefly sketch the steps required for a generalization of eq . and anticipate the final result . in the three - dimensional case
, we found three families of tensorial eigenvalues : one eigenvalue @xmath223 corresponding to an eigenvector in @xmath79 direction , nine eigenvalues corresponding to critical points of @xmath3 via eq . and two eigenvalues introduced by the regularization in eq . .
the same families exist in the general case : we have one eigenvalue @xmath223 with an eigenvector in @xmath79 direction , @xmath224 eigenvalues corresponding to the critical points of @xmath3 in eq . via eq . and @xmath225 eigenvalues introduced by the regularization of the right - hand side . in total
the number of eigenvalues is equal to the degree of the discriminant @xmath226 , see eq . .
in complete analogy to the three - dimensional case one can derive a product formula for the values of @xmath3 evaluated at a subset of its critical points .
it reads @xmath227,\end{aligned}\ ] ] where @xmath135 denotes the solutions of eq .
and @xmath228 is a function which has not been calculated yet .
it is defined by @xmath229}\right]\prod_{j=2 + 3q_c}^{p\cdot 2^{p-1}}\mu_j^{-1},\end{aligned}\ ] ] where the product in the last factor contains all eigenvalues introduced by the regularization via @xmath230 ( cf .
. ref .
@xcite contains a partial argument why @xmath231 should hold .
pending the completion of that argument , we can again draw the conclusion that @xmath232 > 0 \quad\rightarrow\quad \max_{\vartheta}\omega > 0\end{aligned}\ ] ] holds in every physical region of the moduli space .
using the machinery developed above , we now study three important classes of examples in more detail . the eigenvalue problem eq .
is in general difficult to solve explicitly .
often feasible however is the determination of zero eigenvectors , i.e. solving the problem for @xmath233 , which exist if and only if the discriminant of @xmath234 vanishes . in this case , @xmath3 vanishes at the corresponding critical point and it may be of interest to take subleading contributions to the intersection tensor into account .
let @xmath235 denote a zero eigenvector : @xmath236 we now determine the solutions of the perturbed problem in which @xmath237 to leading order in the small parameter @xmath206 . plugging the ansatz @xmath238 into the eigenvector equations and multiplying with @xmath235 we find
is fixed by the condition that the linear system of equations specifying @xmath239 has a solution . ]
@xmath240 expanding eq . in @xmath206
, we obtain @xmath241 and thus for the corresponding @xmath3 @xmath242 as an example , we consider @xmath88 and the simple factorizing volume @xmath243 this model has three vanishing eigenvalues with eigenvectors @xmath244 and @xmath3 therefore vanishes at the corresponding critical points : @xmath245 if we perturb the model as in eq . , we can use eq . and eq . to obtain @xmath246 the next case we are going to study are @xmath88-dimensional models with purely diagonal intersection numbers , i.e. @xmath247 and all other @xmath70 vanish .
in this case @xmath248 is diagonal and it holds @xmath249 we choose orthonormal basis vectors orthogonal to @xmath79 by @xmath250 and @xmath251 where @xmath252 is a normalization constant given by @xmath253 with these choices we have @xmath254 .
if @xmath80 is parameterized as in eq . , the critical points of @xmath134 are given by the solutions of @xmath255 for vanishing @xmath256 , one solution is given by @xmath257 , i.e. @xmath3 has a critical point at @xmath258 .
it can be checked by a direct calculation that choosing @xmath181 different from eq .
as @xmath259 also results in @xmath254 and therefore gives the other two critical points of @xmath3 . in total , we find @xmath260 } = \frac{9}{32}e^{10k}d_{333}^2\frac{\left(d_{111}d_{222}k^1k^2\right)^4}{c_{12}\det g^3 } > 0\displaybreak[0]\\ \omega(\vartheta_2 ) & = \frac{-9e^{-2k}}{(k^2)^3d_{222}\left[d_{111}(k^1)^3+d_{333}(k^3)^3\right ] } = \frac{9}{32}e^{10k}d_{222}^2\frac{\left(d_{111}d_{333}k^1k^3\right)^4}{c_{13}\det g^3 } > 0\displaybreak[0]\\
\omega(\vartheta_3 ) & = \frac{-9e^{-2k}}{(k^1)^3d_{111}\left[d_{222}(k^2)^3+d_{333}(k^3)^3\right ] } = \frac{9}{32}e^{10k}d_{111}^2\frac{\left(d_{222}d_{333}k^2k^3\right)^4}{c_{23}\det g^3 } > 0.\end{aligned}\ ] ] at all three critical points @xmath3 is always positive as long as @xmath96 holds .
another @xmath88-dimensional example which can be treated in more detail is given by a volume factorizing as @xmath261 where @xmath262 is a vector and @xmath263 is a symmetric non - degenerated matrix .
@xmath264 has two vanishing tensorial eigenvalues and by eq .
@xmath3 therefore has to vanish at two of its critical points : @xmath265 to make this more explicit , we can choose coordinates such that @xmath266 then the corresponding zero eigenvectors are easily computed to be @xmath267 where @xmath268 are normalization constants .
we now derive an expression for @xmath269 which depends only on the scalar product of @xmath270 and @xmath271 . for this , choose @xmath268 such that @xmath272 satisfies @xmath273 .
note that @xmath272 are unit vectors only if they are real , because their norm is given by @xmath274 .
then it holds ( due to @xmath275 being zero eigenvectors of @xmath70 ) that @xmath276 this implies ( possibly after changing the orientation ) that @xmath277 where @xmath278 are normalized vectors orthogonal to @xmath79 . according to eq .
the vectors @xmath278 are the extremizers of @xmath3 corresponding to the critical points in eq . .
now we make the ansatz @xmath279 for the eigenvalue problem eq . .
plugging @xmath115 into the eigenvector equations gives @xmath280\nonumber\\ & \quad\quad=\lambda\left(3\alpha^2 k_i+2\alpha\beta\left(\frac{2}{3}k_i+v_{+i}\right)+2\alpha\gamma\left(\frac{2}{3}k_i+v_{-i}\right)+\beta^2\left(\frac{2}{3}v_{+i}+\frac{1}{3}k_i\right)\right.\nonumber\\ & \quad\quad\qquad\left.+\gamma^2\left(\frac{2}{3}v_{-i}+\frac{1}{3}k_i\right)+\frac{2}{3}\beta\gamma\left(\eta k_i+v_{+i}+v_{-i}\right)\vphantom{3\alpha^2}\right),\end{aligned}\ ] ] where we defined @xmath281 by multiplying with @xmath79 , @xmath282 and @xmath283 we obtain the system of equations @xmath284\\ & -2\alpha^2 + 2\alpha\gamma\left(\eta-1\right)\nonumber\\ & \quad= \lambda\left(3\alpha^2+\frac{10}{3}\alpha\beta+\frac{2}{3}\alpha\gamma\left(2 + 3\eta\right)+\beta^2+\frac{1}{3}\gamma\left(\gamma+2\beta\right)\left(2\eta+1\right)\right)\displaybreak[0]\\ & -2\alpha^2 + 2\alpha\beta\left(\eta-1\right)\nonumber\\ & \quad= \lambda\left(3\alpha^2+\frac{10}{3}\alpha\gamma+\frac{2}{3}\alpha\beta\left(2 + 3\eta\right)+\gamma^2+\frac{1}{3}\beta\left(\beta+2\gamma\right)\left(2\eta+1\right)\right).\end{aligned}\ ] ] this system has a solution with a ( potentially ) real @xmath26 , where @xmath285 and @xmath117 and @xmath26 are given by relatively complicated expressions .
fortunately , the @xmath3 corresponding to this eigenvalue simplifies considerably and is given by @xmath286 because @xmath287 it always holds that @xmath288 and therefore @xmath289 we can now compute @xmath290 for @xmath272 defined in eq .
and then use eq . to obtain @xmath291
the quantity @xmath292 turns out to be the first aronhold invariant @xmath293 $ ] , the degree-4 invariant of cubic polynomials in three dimensions we used at the end of section [ sec : product_formula ] to express the discriminant @xmath142 $ ] in terms of simpler invariants . by performing arbitrary rotations to eliminate the restriction in eq . and using the invariance of @xmath294 , @xmath295 and @xmath296 under these rotations , it follows that the general formula has to be @xmath297\,\det d_{ij}\frac{\left(d_ik^i\right)^3}{\det g^3}.\end{aligned}\ ] ] since in the special coordinate system considered above @xmath294 can be written as the square of a real number , it has to be positive in this class of models . using @xmath298
it follows that @xmath299\frac{\det d_{ij}}{d_{jk}k^jk^k}\frac{\left(d_ik^i\right)^2}{\det g^3}.\end{aligned}\ ] ] in particular , if @xmath300 is positive or negative definite , the factor @xmath301 is always positive and @xmath3 is negative .
if @xmath300 is indefinite , the sign of this factor constitutes a simple and direct constraint on the allowed values of @xmath302 .
in this paper we studied constraints on moduli spaces of heterotic string compactifications imposed by the required existence of metastable classical de sitter vacua , assuming that only moduli fields participate in supersymmetry breaking .
we concentrated on three - dimensional moduli spaces and gave the generalization of the two - dimensional result eq . , which has first been derived in @xcite , in eq . .
this equation encodes a rather non - trivial result : if the sign of a degree-12 invariant the discriminant of the calabi - yau intersection tensor is positive , the metastability condition is automatically satisfied on all physically acceptable points on the moduli space .
numerical studies suggest that if the discriminant is negative , metastable de sitter vacua still exist in the generic case , but only for a restricted set of moduli configurations . as briefly discussed in section [ sec : general_case ] , generalizations of eq . seem to exist for arbitrary - dimensional moduli spaces , raising the question of the existence of a more intuitive interpretation of the discriminant .
we also studied specific examples of three - dimensional moduli spaces . for moduli spaces with dimension @xmath2 ,
the metastability analysis is difficult to carry out explicitly and the result in general depends on non - topological properties of the calabi - yau , in this case its khler structure ( cf .
eq . and eq . and the numerical result in fig .
[ fig : heterotic ] ) .
this complication already appears in the three - dimensional case . while the reduction of the problem from eq . to eq .
essentially reduces the problem to the task of finding the roots of a cubic polynomial , the extraction of meaningful results has not yet been successful in the general case . on the other hand
, the class of calabi - yaus studied in section [ sec : almost_fact_3 ] constitutes a promising candidate for further studies : it naturally generalizes the class of factorizable models , i.e. models with a volume of the form @xmath303 where @xmath304 run from @xmath305 to @xmath73 , while not suffering from the fact that factorizable models do not allow metastable de sitter vacua without invoking higher - order corrections or additional tree - level contributions@xcite .
i would like to thank laura covi for helpful discussions and comments .
in this appendix we demonstrate that for @xmath88-dimensional moduli spaces of heterotic string models the phases @xmath306 and @xmath307 in eq . can safely be set to zero in the metastability analysis . to see this , we parameterize @xmath80 as in eq . and write @xmath3 as @xmath308,\end{aligned}\ ] ] where @xmath309 by swapping the sign of @xmath310 , the term @xmath311 can always be made a positive contribution to @xmath3 , which is maximal if @xmath312 .
the term proportional to @xmath313 can potentially give a negative contribution , which can be reduced ( or even turned to a positive one ) if the complex phases do not vanish .
hence , we find that the global maximum of @xmath3 in the three - dimensional case @xmath88 can only have a non - vanishing ( non - global ) phase @xmath314 if @xmath315 the quantity @xmath256 changes sign if the vectors @xmath181 and @xmath182 are rotated into each other by @xmath316 , implying that we can choose a basis @xmath317 such that @xmath254 . in this basis , eq . reads @xmath318 this is only fulfilled if @xmath319 in particular , it has to hold that @xmath320 implying that @xmath321 .
thus , if the global maximum of @xmath3 is attained at @xmath322 , there will always be another critical point of @xmath3 with @xmath312 at which @xmath3 is still positive .
finally , a global phase @xmath323 always drops out of @xmath3 , proving the claim .
after restricting to @xmath88 and setting the complex phases to zero , we found four critical points of @xmath3 as a function of @xmath324 . we now show that , as has been claimed in section [ sec:3_dim ] , that one of these , namely the one given by eq .
, can be discarded a priori in the search for the global maximum of @xmath3 . as in the last section
, we use the freedom in choosing @xmath181 and @xmath182 to set @xmath254 . then by explicitly solving eq . for the critical point @xmath325 and plugging the result back into @xmath3
we find @xmath326 = -1+\frac{3}{2}\frac{d_{112}^2\left(d_{112}+d_{222}\right)^2+d_{111}^2d_{222}^2}{d_{111}^2+\left(d_{112}+d_{222}\right)^2}.\end{aligned}\ ] ] to prove that this can never constitute the global maximum of @xmath3 , we have to show that @xmath327}d_{nnn}^2(\vartheta)\geq\frac{d_{112}^2\left(d_{112}+d_{222}\right)^2+d_{111}^2d_{222}^2}{d_{111}^2+\left(d_{112}+d_{222}\right)^2},\end{aligned}\ ] ] where @xmath328 we abbreviate @xmath329 , @xmath330 and @xmath331 . if @xmath332 , setting @xmath257 shows the claim .
if @xmath333 , we can set @xmath334 and are done .
let us thus assume that @xmath335 and @xmath336 .
then @xmath337 and the claim follows if there is an @xmath338 such that @xmath339 an extremum in @xmath338 of the left - hand side of eq . must satisfy @xmath340 this equation is in fact only quadratic in @xmath341 and is solved by @xmath342 taking the ` @xmath343 ' solution if @xmath344 and the ` @xmath345 ' solution if @xmath346 we see that ( by concavity of the square root ) @xmath347 so @xmath348 the next step is to show that @xmath349 grows monotonically for @xmath350 .
this can be seen from @xmath351 because @xmath352 with @xmath353 and @xmath354 this finally gives for @xmath350 @xmath355 the calculation for @xmath346 is analogous .
alternatively , the claim follows by substituting @xmath356 in eq . .
k. becker , m. becker , m. haack and j. louis , _ supersymmetry breaking and alpha-corrections to flux induced potentials _ , jhep * 0206 * , 060 ( 2002 ) [ arxiv : hep - th/0204254 ]
. v. balasubramanian and p. berglund , _ stringy corrections to kaehler potentials , susy breaking , and the cosmological constant problem _ , jhep * 0411 * ( 2004 ) 085 [ arxiv : hep - th/0408054 ] .
s. l. parameswaran and a. westphal , _ de sitter string vacua from perturbative kaehler corrections and consistent d - terms _ , jhep * 0610 * ( 2006 ) 079 [ arxiv : hep - th/0602253 ] .
e. palti , g. tasinato and j. ward , _ weakly - coupled iia flux compactifications _ , jhep * 0806 * ( 2008 ) 084 [ arxiv:0804.1248 ] .
m. berg , m. haack and e. pajer , _ jumping through loops : on soft terms from large volume compactifications _
, jhep * 0709 * ( 2007 ) 031 [ arxiv:0704.0737 ] .
f. f. gautason , d. junghans and m. zagermann , _ on cosmological constants from alpha-corrections _ , jhep * 1206 * ( 2012 ) 029 [ arxiv:1204.0807 ] .
s. kachru , r. kallosh , a. linde and s. p. trivedi , _ de sitter vacua in string theory _ , phys
. rev .
d * 68 * ( 2003 ) 046005 [ arxiv : hep - th/0301240 ] .
e. silverstein , _ simple de sitter solutions _ , phys .
d * 77 * ( 2008 ) 106006 [ arxiv:0712.1196 ] .
s. krippendorf , f. quevedo , _ metastable susy breaking , de sitter moduli stabilisation and khler moduli inflation _
, jhep * 0911 * ( 2009 ) 039 [ arxiv:0901.0683 ] .
s. s. haque , g. shiu , b. underwood and t. van riet , _ minimal simple de sitter solutions _ , phys . rev .
d * 79 * ( 2009 ) 086005 [ arxiv:0810.5328 ] .
u. h. danielsson , s. s. haque , g. shiu and t. van riet , _ towards classical de sitter solutions in string theory _
, jhep * 0909 * ( 2009 ) 114 [ arxiv:0907.2041 ] .
m. gomez - reino and c. a. scrucca , _ locally stable non - supersymmetric minkowski vacua in supergravity _ , jhep * 0605 * ( 2006 ) 015 [ arxiv : hep - th/0602246 ] .
r. brustein and s. p.
de alwis , _ moduli potentials in string compactifications with fluxes : mapping the discretuum _ , phys .
d * 69 * ( 2004 ) 126006 [ arxiv : hep - th/0402088 ] .
l. covi , m. gomez - reino , c. gross , g. a. palma and c. a. scrucca , _ constructing de sitter vacua in no - scale string models without uplifting _ , jhep * 0903 * ( 2009 ) 146 [ arxiv:0812.3864 ] .
l. covi , m. gomez - reino , c. gross , j. louis , g. a. palma and c. a. scrucca , _ de sitter vacua in no - scale supergravities and calabi - yau string models _ ,
jhep * 0806 * ( 2008 ) 057 [ arxiv:0804.1073 ] . c. gross , _ de sitter vacua and inflation in no - scale string models _ ,
dissertation , universitt hamburg , 2009 , desy report : desy - thesis-2009 - 029 .
m. gomez - reino and c. a. scrucca , _ metastable supergravity vacua with f and d supersymmetry breaking _ , jhep * 0708 * ( 2007 ) 091 [ arxiv:0706.2785 ] .
m. gomez - reino , j. louis and c. a. scrucca , _ no metastable de sitter vacua in n=2 supergravity with only hypermultiplets _
, jhep * 0902 * ( 2009 ) 003 [ arxiv:0812.0884 ] .
p. candelas , g. t. horowitz and a. strominger , _ vacuum configurations for superstrings _ , nucl .
b * 258 * ( 1985 ) 46 .
l. aparicio , d. g. cerdeo and l. e. ibez , _ a 119 - 125 gev higgs from a string derived slice of the cmssm _
, jhep * 1204 * ( 2012 ) 126 [ arxiv:1202.0822 ] .
d. farquet and c. a. scrucca , _ scalar geometry and masses in calabi - yau string models _
, jhep * 1209 * ( 2012 ) 025 [ arxiv:1205.5728 ] .
p. candelas and x. de la ossa , _ moduli space of calabi - yau manifolds _ , nucl . phys .
b * 355 * ( 1991 ) 455 .
b. r. greene and m. r. plesser , _ duality in calabi - yau moduli space _
b * 338 * ( 1990 ) 15 .
d. rathlev , _ de sitter vacua in no - scale supergravity models _ , master s thesis , universitt gttingen , unpublished . | we study the possibility of obtaining metastable de sitter vacua of heterotic string theory compactified on a calabi - yau threefold which are classical and simple in the khler moduli sector of the theory . for this , we exploit a known necessary condition on the khler potential in @xmath0-supergravity , which we , under the assumption that only moduli fields contribute to supersymmetry breaking , express in terms of a tensorial eigenvalue problem for the calabi - yau triple intersection tensor . for three - dimensional moduli spaces we are able to identify the discriminant of the calabi - yau intersection tensor in the analysis , generalizing a known result for two - dimensional moduli spaces .
we also discuss explicit examples and possible generalizations .
0.4 cm * constraints from de sitter metastability in heterotic string compactifications * 0.7 cm dirk rathlev 0.5 cm _ institut fr theoretische physik , universitt zrich , + ch-8057 zrich , switzerland + _ 0.8 cm |
dynamical systems with invariant manifolds in phase space have been the subject of ongoing research in recent years .
many authors have considered the case of two or more coupled identical systems that are chaotic .
on invariant manifolds the subsystems display identical or synchronized motion , and the manifold s transverse stability is investigated @xcite .
an alternative approach is based on the observation that any rotationally invariant system of identical interacting particles possesses low dimensional invariant manifolds in classical phase space @xcite . on such manifolds , the classical motion displays largely collective behavior and decouples from more complex single - particle behavior .
the importance of a given invariant manifold depends crucially on its stability properties .
if the manifold under consideration is sufficiently stable in transverse directions , the quantum system may exhibit wave function scarring @xcite or display a strong revival for wave - packets localized to the vicinity of the manifold @xcite .
these findings may be directly associated with the slow decay of collective motion due to the coupling between collective and single - particle motion . in this paper
we propose suitably adapted coordinate systems that separate collective and single - particle motion on the invariant manifolds mentioned above .
such coordinates clarify the separation of collective and single particle motion and may be useful in several applications .
we have in mind ( i ) the problem of damping and dissipation of collective excitations and the interplay of collective and chaotic motion in atomic nuclei @xcite , which is often addressed in the framework of single - particle motion in a time dependent mean field ; ( ii ) multi particle fragmentation of atoms at threshold which evolves over highly symmetric configurations corresponding to invariant manifolds @xcite ; ( iii ) the structural stability of invariant manifolds @xcite .
there is a traditional way to introduce collective and single - particle coordinates in interacting many - body systems .
aiming at the description of nuclear vibrations and rotations , zickendraht @xcite introduced a system of collective coordinates in a self - bound many - body system .
three of these coordinates describe the center of mass motion , and six collective degrees of freedom govern the dynamics of the inertia ellipsoid .
the remaining coordinates are of single - particle nature .
we shall establish the relation of coordinates of invariant manifolds to those defined by zickendraht .
furthermore we shall show that more complicated collective motion , e.g. shearing modes can be described this article is divided as follows . in the next section
we introduce suitable coordinate systems for interacting many - body systems with invariant manifolds .
we give a construction recipe and present a detailed example calculation . as an application we give a potential expansion around an invariant manifold and discuss stability properties . in section [ zick ]
we make a connection with the zickendraht coordinates .
we present examples where the motion of the inertia ellipsoid corresponds to the motion on an invariant manifold . for such initial conditions the traditional collective motion decouples completely from the single - particle degrees of freedom .
we also find that collective coordinates as defined here are capable of other types of motion .
therefore we finally discuss how motion on or near such invariant manifolds could be interpreted as collective motion of a system .
in this section we present a transformation from cartesian single particle coordinates in position and momentum space to cartesian coordinates that are adapted to invariant manifolds .
the new coordinates consist of `` collective coordinates '' that govern the motion on the invariant manifold and of coordinates transversal to this manifold that represent the single - particle aspects .
consider rotationally invariant systems of @xmath0 identical particles in @xmath1 spatial dimensions ( @xmath2 or @xmath3 ) .
the hamiltonian is invariant under both , the action of the rotation group o(@xmath1 ) and the group of permutations s@xmath4 .
one may now take a finite subgroup @xmath5 o(@xmath1 ) with elements @xmath6 and properly chosen permutations @xmath7 such that [ mandef ] g p^-1(g ) ( , ) = ( , ) , g ( p_1, ,p_nd ) , ( q_1, ,q_nd ) for points @xmath8 on some invariant submanifold of phase space .
on such a manifold , the action of certain rotations @xmath6 can be canceled by permutations .
these permutations clearly form a subgroup isomorphic to @xmath9 .
[ fig1 ] shows a configuration of four particles in two spatial dimensions that corresponds to a point on an invariant manifold .
the operations of elements from the discrete symmetry group @xmath10 can be undone by suitable permutations of particles .
this leads to a collective motion with two degrees of freedom which we shall identify with vibrations .
[ fig2 ] shows two spatial configurations of eight ( [ fig2]a ) and six ( [ fig2]b ) particles , respectively which display a @xmath11 symmetry .
if initial momenta display the same symmetry the motion on the invariant manifold will have two degrees of freedom . for eight particles
the radii of the two circles will oscillate synchronously , and the two circles will vibrate against each other .
for the six particles we will have a vibration of the radius of the circle and of the two particles along the vertical axis .
we may choose initial momenta to reduce the symmetry group to @xmath12 which will allow rotations around the vertical axis and thus add an additional degree of freedom .
for eight particles we could alternatively choose initial conditions that are limited to a @xmath13 symmetry . besides the vibrations discussed above this would allow for a shearing motion of the two circles thus yielding again three degrees of freedom .
we could also reduce the fourfold rotation axis to a twofold one and obtain @xmath14 , @xmath15 or @xmath16 as remaining symmetry groups yielding more collective degrees of freedom . adding
two particles symmetrically onto the principal axis of rotation would also increase the number of degrees of freedom by one .
other reductions of symmetry will yield different invariant manifolds with varying degrees of freedom .
we will see this exemplified by explicit construction of coordinates .
we may use the definition ( [ mandef ] ) directly for the construction of coordinate systems where invariant manifolds correspond to coordinate axis or planes , i.e. non collective coordinates vanish for motion on the invariant manifold . to this purpose
we consider the many - body system in cartesian coordinates in momentum and position space . in
what follows we will introduce orthogonal transformations in configuration space only ; momenta will be subject to the same transformation . in a cartesian coordinate system
each element @xmath17 and each permutation @xmath7 can be represented by an orthogonal matrix @xmath18 and @xmath19 of dimension @xmath20 .
it is clear that the products @xmath21 form a matrix group @xmath22 that acts onto position and momentum space , respectively .
the construction of the coordinate system is now straightforward .
every vector @xmath23 and @xmath24 may be expanded in basis vectors of the irreducible representations ( irs ) of @xmath22 by means of projectors @xcite .
[ projec ] _ = _ g _ g^()*m*_g*p*^t_g . here
@xmath25 denotes the character of @xmath6 in the @xmath26th ir .
similar formulae hold for momentum space .
the projection onto the identical ir defines the invariant manifold .
note that the identical representation is one - dimensional while the invariant manifolds of interest typically have higher dimensionality .
we can find independent vectors on the manifold by projecting from different vectors , but in practice the construction of the independent vectors seems to be unproblematic as we shall see in the example .
a comment on the rotation symmetry is in order . like
any cartesian coordinates , the coordinates introduced in this article do not explicitly reflect the invariance under rotations .
acting on an invariant manifold , rotations generate a continuous family of equivalent manifolds .
our coordinates , however , single out one particular manifold .
for quantum systems , the rotation operator may easily be constructed and used for projection onto subspaces of definite angular momentum .
we now illustrate the proposed construction explicitly for four particles in two dimensions and a quartic potential , considering the invariant manifold shown in fig .
we shall also expand the potential near the invariant manifold to second order in the transversal coordinates .
the invariant manifold is defined by those points which are invariant under @xmath27 , where @xmath28 denotes the identity , @xmath29 a permutation of particles as indicated , @xmath30 a reflection at the axis indicated , and @xmath31 a rotation about @xmath32 .
thus , @xmath33 with four irs labeled by @xmath34 @xcite .
let @xmath35 denote a coordinate vector in position space ( @xmath36 denote the coordinates of the @xmath37th particle ) .
we have e & = & ( x_1,x_2,x_3,x_4,y_1,y_2,y_3,y_4 ) , + _ x p_(12)(34 ) & = & ( x_2,x_1,x_4,x_3,-y_2,-y_1,-y_4,-y_3 ) + c_2p_(13)(24 ) & = & ( -x_3,-x_4,-x_1,-x_2,-y_3,-y_4,-y_1,-y_2 ) + _ y p_(14)(23 ) & = & ( -x_4,-x_3,-x_2,-x_1,y_4,y_3,y_2,y_1 ) .
using the character table of @xmath16 @xcite and the projectors ( [ projec ] ) one constructs the following basis vectors corresponding to the ir labeled by a_1 & : & e_1=(1,1,-1,-1,0,0,0,0)/2 , e_2=(0,0,0,0,1,-1,-1,1)/2 , + b_1 & : & e_3=(1,1,1,1,0,0,0,0)/2 , e_4=(0,0,0,0,1,-1,1,-1)/2 , + a_2 & : & e_5=(1,-1,-1,1,0,0,0,0)/2 , e_6=(0,0,0,0,1,1,-1,-1)/2 , + b_2 & : & e_7=(1,-1,1,-1,0,0,0,0)/2 , e_8=(0,0,0,0,1,1,1,1)/2 .
the vectors associated with the identical ir @xmath38 span the two - dimensional invariant manifold and the vectors associated with the irs @xmath39 span the transverse directions .
we now present the orthogonal transformation that transforms the single particle coordinates @xmath24 into the coordinates adapted to the invariant manifold . in our example @xmath40 and @xmath41-components do not mix and we have [ trans ] = 12 , = 12 . to illustrate the example and to further demonstrate the usefulness of the newly introduced coordinate system we want to consider the the interacting four - body system with hamiltonian [ ham ] h=_i=1 ^ 4((p_x_i^2+p_y_i^2)/2 + 16(x_i^2+y_i^2)^2 ) -_i < j .
this hamiltonian has been studied previously @xcite .
in particular , the stability of the invariant manifold displayed in fig .
[ fig1 ] has been studied by computing the full phase space monodromy matrix of several periodic orbits that are inside the invariant manifold .
it was found that several orbits are linearly stable in transverse directions or possess rather small stability exponents .
qualitatively , this may also be understood by studying the hamiltonian ( [ ham ] ) close to the invariant manifold .
we therefore use the transformation ( [ trans ] ) and expand the potential of hamiltonian ( [ ham ] ) to second order in the transverse directions labeled by ( @xmath42 ) while keeping the full dependence of the coordinates ( @xmath43 ) inside the invariant manifold .
one obtains the quadratic form @xmath44 where = .
a diagonalization of @xmath45 yields the eigenvalues _ 1,2 & = & 10x^2 + 14y^22 , +
_ 3,4 & = & 14x^2 + 10y^22 , + _ 5,6 & = & 6(x^2+y^2 ) 2.all eigenvalues are non - negative and vanish at the origin @xmath46 .
thus , instability may occur only in its vicinity . though the expansion of a potential around an invariant manifold is no substitute for the computations of lyapunov exponents or monodromy matrices ,
it is a first step when estimating stability properties of such manifolds .
almost thirty years ago zickendraht @xcite introduced a set of collective coordinates to describe nuclear vibrations and rotations , as well as their coupling with single particle motion .
we shall discuss to what extent these coordinates correspond to the ones we introduced in the previous sections . on one hand this will allow to identify certain vibrational modes of a many - body system with invariant manifolds . on the other hand we shall also see that our procedure proposes collective movements that are not of the type described easily in zickendraht s coordinates .
following zickendraht @xcite we write the coordinates @xmath47 of the @xmath48 particle in the center of mass system as [ zickone ] _
i = s_i1+s_i2+s_i3,i=1, ,n where the @xmath49 span the inertia ellipsoid and @xmath50 are non - collective coordinates which for simplicity we shall call single - particle coordinates .
the newly introduced coordinates @xmath49 and @xmath51 are not independent .
the constraints are _ i_j & = & y_i y_j _ ij , i , j=1,2,3 + _ i=0^n s_ij & = & 0,j=1,2,3 + _
i=0^n s_ijs_ik&=&_jk , j , k=1,2,3.the first six equations ensure the orthogonality and normalization of the principal axis of the inertia ellipsoid whereas the next three equations fix the origin at the center of mass system .
the last six equations are orthogonality relations of the single - particle coordinates . in the center of mass system , one may therefore characterize the @xmath0-body system by its inertia ellipsoid ( e.g. three euler angles of the principle axis and three moments of inertia ) and @xmath52 single particle coordinates .
the moments of inertia @xmath53 are related to the coordinates @xmath54 by [ inert ] i_1=m(y_2 ^ 2+y_3 ^ 2),i_2=m(y_1 ^ 2+y_3 ^ 2),i_3=m(y_1 ^ 2+y_2 ^ 2 ) , where @xmath55 denotes the mass of the particles .
it is interesting to determine those configurations , where the motion of the many - body system may be described in terms of the collective coordinates @xmath54 only . while such motion would be restricted to
_ some _ invariant manifold in phase space it would not obviously be one of those defined by eq .
( [ mandef ] ) .
we may however determine invariant manifolds ( [ mandef ] ) such that the motion on the manifold changes only the inertia ellipsoid of the system and hence may be described entirely by zickendraht s collective coordinates @xmath54 .
two necessary conditions for this a situation are easily stated .
first , the number of coordinates on such invariant manifold may not exceed six in the general case and three in the case of pure vibrations .
second , every motion on such an invariant manifold has to change the inertia ellipsoid of the many - body system . for simplicity
let us start with the a system of four particles in two spatial dimensions and the invariant manifold displayed in fig .
[ fig1 ] , i.e. _ 1=,_2=,_3=,_4=,and the momenta are chosen by replacing @xmath56 .
computation of the moments of inertia yield the collective zickendraht coordinates @xmath57 , @xmath58 . on the invariant manifold
the remaining coordinates are given by @xmath59 .
this shows that every motion on the invariant manifold only changes the moments of inertia and therefore decouples from the single - particle motion .
we next consider the example of an eight - body system in three dimensions .
let [ conf ] _
1=,_2=,_3=,_4=,_4+i=_i(z - z ) denote a configuration restricted to the invariant manifold displayed in fig .
[ fig2 ] ( a ) with @xmath12 symmetry .
( the momenta are chosen by replacing @xmath60 in eq.([conf ] ) . )
the moments of inertia are @xmath61 and yield collective coordinates ( [ inert ] ) @xmath62 . since the inertia ellipsoid is symmetric we have a freedom in choosing two of its principle axis . using _
1=2,_2=2,_3=.one obtains constant single - particles coordinates @xmath63 for the motion on the invariant manifold .
thus , the single - particle motion decouples from the collective motion on the invariant manifold .
similar results hold for the six particle configuration displayed in fig .
[ fig2 ] .
it is also instructive to consider one counterexample .
the configuration _ 1=,_2=,_3=,_4= , + _
5=,_6=,_7=,_8=,displays @xmath13 symmetry and differs from configuration ( [ conf ] ) by a shearing motion . like in the previous example ,
the moments of inertia are given by @xmath61 and the ellipsoid of inertia is symmetric .
however , no choice of the principal axis allows to fulfill eqs .
( [ zickone ] ) with _ constant _ single - particle coordinates @xmath51 .
therefore , single - particle degrees of freedom depend on collective degrees of freedom and a decoupling does not exist using zickendraht s coordinate system .
a decoupling is obtained by using the coordinates introduced in this work .
however , the collective motion on the appropriate invariant manifold does not correspond to pure vibrations or rotations of the inertia ellipsoid .
these findings are interesting e.g. in relation with with the magnetic dipole mode in nuclei @xcite since this type of collective behavior is associated with a shearing motion .
we constructed an orthogonal transformation that maps the cartesian single particle coordinates of a many - body system to a new cartesian coordinate system that distinguishes collective and single - particle motion .
the collective degrees of freedom govern the motion that is restricted to a low - dimensional invariant manifold and are decoupled from single - particle degrees of freedom on this manifold .
we have demonstrated that there are several configurations of few - body systems , where the motion on the invariant manifold corresponds to a vibration or rotation and may be described in terms of zickendraht s collective coordinates , but differs when the collective motion goes beyond that .
these results are independent of the details of the hamiltonian of the @xmath0-body system , and are entirely determined by rotational and permutational symmetry . using the results of this article as well as those of refs .
@xcite we can draw the following picture : first it is possible that an invariant manifold is spanned exactly by the vibrational and rotational modes of a few - body system ; second such manifolds may be stable or have small instability exponents in transversal directions ; third the revival probabilities of wave packets launched on such manifolds are large ; last , as a conclusion of these points we may have a collective motion near the manifold whose damping is characterized by the decay rate in transversal direction .
we also found that there may be other collective motions ; this was displayed in an example of shearing motion , but there can be others such as breathing modes etc .
the coincidence of zickendraht coordinates with our collective coordinates depends on particle numbers ; typically they do not span an invariant manifold .
this confirms the well - known fact that in general the collective motion in these coordinates does not separate rigorously , but only in some adiabatic approximation .
as we found more general invariant configurations which in turn induce collective coordinates we may hope that these are useful for approximate considerations for larger particle numbers that do include the corresponding invariant manifold in a non - trivial fashion .
the construction of appropriate coordinates is an open problem .
t.p . acknowledges the warm hospitality at the centro internacional de ciencias , cuernavaca , mexico , where part of this work has been done .
this work was partially supported by dept . of energy ( usa ) , by dgapa ( unam ) and by conacyt ( mexico ) .
j. f. heagy , t. l. caroll , and l. m. pecora , ( 1994 ) 1874 r. brown , n. f. rulkov , and n. b. tufillaro , ( 1994 ) 4488 n. f. rulkov , m. m. shushchick , l. s. tsimring , and h. d. i. abarbanel , ( 1995 ) 980 d. beigie , chaos , solitones and fractals * 5 * ( 1995 ) 177 d. j. gauthier and j. c. bienfang , ( 1996 ) 1751 s. c. venkataramani , t. m. antonsen , e. ott , and j. c. sommerer , physica d * 96 * ( 1996 ) 66 j. yang , g. hu , and j. xiao , ( 1998 ) 496 t. papenbrock and t. h. seligman , a * 218 * ( 1996 ) 229 d. wintgen , k. richter , and g. tanner , chaos * 2 * ( 1992 ) 19 t. prosen , a * 233 * ( 1997 ) 332 t. papenbrock and t. prosen , ( 2000 ) 262 t. papenbrock , t. h. seligman , and h. a. weidenmller , ( 1998 ) 3057 t. guhr , h. a. weidenmller , ann
* 193 * ( 1989 ) 472 j. blocki , f. brut , t. srokowski , and w. j. swiatecki , nucl .
phys . a * 545 * ( 1992 ) 551c w. d. heiss , r. g. nazmitdinov , and s. radu , ( 1994 ) 2351 s. drozdz , s. nishizaki , and j. wambach , ( 1994 ) 2839 w. bauer , d. mcgrew , v. zelevinski , and p. schuck , ( 1994 ) 3771 v. r. manfredi , l. salasnich , int .
e * 4 * ( 1995 ) 625 v. zelevinski , ann .
nucl . part .
* 46 * ( 1996 ) 237 t. papenbrock , ( 2000 ) 034602 t. pattard and j. m. rost , ( 1998 ) 5081 l. benet , c. jung , t. papenbrock , and t. h. seligman , physica d * 131 * ( 1999 ) 254 w. zickendraht , j. math
* 12 * ( 1970 ) 1663 m. hammermesh , _ group theory and its application to physical problems _ , dover publications , n.y .
, 1989 b. eckhardt , g. hose , and e. pollak , ( 1989 ) 3776 t. guhr , h. diesener , a. richter , c. w. de jager , h. de vries , and p. k. a. de witts - huberts , z. phys .
a * 336 * ( 1990 ) 159 | we introduce suitable coordinate systems for interacting many - body systems with invariant manifolds . these are cartesian in coordinate and momentum space and chosen such that several components are identically zero for motion on the invariant manifold . in this sense
these coordinates are collective .
we make a connection to zickendraht s collective coordinates and present certain configurations of few - body systems where rotations and vibrations decouple from single - particle motion .
these configurations do not depend on details of the interaction .
# 1eq .
( [ # 1 ] ) |
multiple - access channels and two - way channels are two of the earliest channels that were considered in the literature . the multiple - access channel capacity region was determined in @xcite .
the two - way channel was initially examined by shannon , @xcite , where he found inner and outer bounds for the general two - way channel , and determined the capacity region for some special cases . in @xcite
, it was shown that the inner bound found by shannon was not tight in general .
the capacity region of the gaussian two - way channel was found by han in @xcite .
a related , somewhat more general case called two - user channels was studied in @xcite . for a comprehensive review of these channels ,
the reader is referred to @xcite .
a rigorous analysis of information theoretic secrecy was first given by shannon in @xcite . in this work
, shannon showed that to achieve _ perfect secrecy _ in communications , which is equivalent to providing no information to an enemy cryptanalyst , the conditional probability of the _ cryptogram given a message _ must be independent of the actual transmitted message . in other words , the _ a posteriori _ probability of a message must be equivalent to its _ a priori _ probability . in @xcite
, wyner applied this concept to the discrete memoryless channel .
he defined the wire - tap channel , where there is a wire - tapper who has access to a degraded version of the intended receiver s signal . using the normalized conditional entropy @xmath0 of the transmitted message given the received signal at the wire - tapper as the secrecy measure
, he found the region of all possible @xmath1 pairs , and the existence of a _ secrecy capacity _ , @xmath2 , the rate up to which it is possible to limit the rate of information transmitted to the wire - tapper to arbitrarily small values . in @xcite , it was shown that for wyner s wire - tap channel , it is possible to send several low - rate messages , each completely protected from the wire - tapper individually , and use the channel at close to capacity .
however , if any of the messages are available to the wire - tapper , the secrecy of the rest may also be compromised .
reference @xcite extended wyner s results in @xcite and carleial and hellman s results in @xcite to gaussian channels . the seminal work by csiszr and krner , @xcite , generalized wyner s results to less noisy " and more capable " channels .
furthermore , it examined sending common information to both the receiver and the wire - tapper , while maintaining the secrecy of some private information that is communicated to the intended receiver only .
reference @xcite suggested that the secrecy constraint developed by wyner needed to be strengthened , since it constrains the rate of information leaked to the wire - tapper , rather than the total information , and the information of interest might be in this small amount .
it was then shown that the results of @xcite can be extended to
strong " secrecy constraints for discrete channels , where the limit is on the total leaked information rather than just the rate , with no loss in achievable rates , @xcite . in the past two decades ,
common randomness has emerged as a valuable resource for secret key generation , @xcite . in @xcite , it was shown that the existence of a
public " feedback channel can enable the two parties to be able to generate a secret key even when the wire - tap capacity is zero .
references @xcite and @xcite examined the secret key capacity and _ common randomness _
capacity , for several channels .
these results also benefit from @xcite to provide strong " secret key capacities .
maurer also examined the case of active adversaries , where the wire - tapper has read / write access to the channel in @xcite @xcite .
the secret key generation problem was investigated from a multi - party point of view in @xcite and @xcite .
notably , csiszr and narayan considered the case of multiple terminals where a number of terminals try to distill a secret key and a subset of these terminals can act as helper terminals to the rest in @xcite , @xcite .
recently , several new models have emerged , examining secrecy for parallel channels @xcite , relay channels @xcite , and fading channels @xcite . fading and parallel channels were examined together in @xcite .
broadcast and interference channels with confidential messages were considered in @xcite .
references @xcite examined the multiple access channel with confidential messages where two transmitters try to keep their messages secret from each other while communicating with a common receiver . in @xcite , an achievable region
was found in general , and the capacity region was found for some special cases .
mimo channels were considered in @xcite . in @xcite , we investigated multiple access channels where transmitters communicate with an intended receiver in the presence of an external wire - tapper from whom the messages must be kept confidential . in @xcite , we considered the case where the wire - tapper gets a degraded version of a gmac signal , and defined two separate secrecy measures extending wyner s measure to multi - user channels to reflect the level of trust the network may have in each node .
achievable rate regions were found for different secrecy constraints , and it was shown that the secrecy sum - capacity can be achieved using gaussian inputs and stochastic encoders .
in addition , tdma was shown to also achieve the secrecy sum - capacity . in this paper
, we consider the general gaussian multiple access wire - tap channel ( ggmac - wt ) and the gaussian two - way wire - tap channel ( gtw - wt ) , both of which are of interest in wireless communications as they correspond to the case where a single physical channel is utilized by multiple transmitters , such as in an ad - hoc network .
we consider an external _ eavesdropper _ that receives the transmitters signals through a general gaussian multiple access channel ( ggmac ) in both system models .
we utilize a suitable secrecy constraint which is the normalized conditional entropy of the transmitted secret messages given the eavesdropper s signal , corresponding to the collective secrecy " constraints used in @xcite .
we show that satisfying this constraint implies the secrecy of the messages for all users . in both scenarios ,
transmitters are assumed to have one secret and one open message to transmit .
this is different from @xcite in that the secrecy rates are not constrained to be at least a fixed portion of the overall rates .
we find an achievable _ secrecy rate region _
, where users can communicate with arbitrarily small probability of error with the intended receiver under _ perfect secrecy _ from the eavesdropper , which corresponds to the result of @xcite for the degraded case .
we note that , in accordance with the recent literature , when we use the term perfect secrecy , we are referring to weak " secrecy , where the _ rate _ of information leaked to the adversary is limited .
as such , this can be thought of as almost perfect secrecy " .
we also find the sum - rate maximizing power allocations for the general case , which is more interesting from a practical point of view .
it is seen that as long as the users are not _ single - user decodable _ at the eavesdropper , a secrecy - rate trade off is possible between the users .
next , we show that a non - transmitting user can help increase the secrecy capacity for a transmitting user by effectively jamming " the eavesdropper , and even enable secret communications that would not be possible in a single - user scenario .
we term this new scheme _
cooperative jamming_. the gtw - wt is shown to be especially useful for secret communications , as the multiple - access nature of the channel hurts the eavesdropper without affecting the communication rate .
this is due to the fact that the transmitted messages of each user essentially help hide the other user s secret messages , and reduce the extra randomness needed in wire - tap channels to confuse the eavesdropper .
the rest of the paper is organized as follows : section [ sec : system ] describes the system model for the ggmac - wt and gtw - wt and the problem statement .
section [ sec : ach ] describes the general achievable rates for the ggmac - wt and gtw - wt .
sections [ sec : summax ] and [ sec : jam ] give the sum - secrecy rate maximizing power allocations , and the achievable rates with cooperative jamming . section [ sec : results ] gives our numerical results followed by our conclusions and future work in section [ sec : conclusion ] .
we consider @xmath3 users communicating in the presence of an eavesdropper who has the same capabilities .
each transmitter @xmath4 has two messages , @xmath5 which is secret and @xmath6 which is open , from two sets of equally likely messages @xmath7 , @xmath8 . let @xmath9 , @xmath10 , @xmath11 , @xmath12 , and @xmath13 .
the messages are encoded using @xmath14 codes into @xmath15 , where @xmath16 .
the encoded messages @xmath17 are then transmitted .
we assume the channel parameters are universally known , and that the eavesdropper also has knowledge of the codebooks and the coding scheme .
in other words , there is no shared secret .
the two channels we consider in this paper are described next .
this is a scenario where the users communicate with a common base station in the presence of an eavesdropper , where both channels are modeled as gaussian multiple - access channels as shown in figure [ fig : gmacwt2 ] .
the intended receiver and the wire - tapper receive @xmath18 and @xmath19 , respectively .
the receiver decodes @xmath20 to get an estimate of the transmitted messages , @xmath21 .
we would like to communicate with the receiver with arbitrarily low probability of error , while keeping the wire - tapper ( eavesdropper ) ignorant of the secret messages , @xmath22 .
the signals at the intended receiver and the wire - tapper are given by [ eqn : mac ] @xmath23 where @xmath24 are the awgn , @xmath25 is the transmitted codeword of user @xmath26 , and @xmath27 are the channel gains of user @xmath26 to the intended receiver ( _ main _ channel , m ) , and the eavesdropper ( _ wire - tap _ channel , w ) , respectively .
each component of @xmath28 and @xmath29 .
we also assume the following transmit power constraints : @xmath30 similar to the scaling transformation to obtain the standard form of the interference channel , @xcite , we can represent any gmac - wt by an equivalent standard form , @xcite : [ eqn : macstd ] @xmath31 where , for each @xmath26 , * the codewords are scaled to get @xmath32 ; * the new power constraints are @xmath33 ; * the wiretapper s new channel gains are @xmath34 ; * the noises are normalized to get @xmath35 and @xmath36 . we can show that the eavesdropper gets a stochastically degraded version of the receiver s signal if @xmath37 .
we considered this special case in @xcite . in this scenario ,
two transmitter / receiver pairs communicate with each other over a common channel .
each receiver @xmath38 gets @xmath39 and the eavesdropper gets @xmath19 .
receiver @xmath26 decodes @xmath40 to get an estimate of the transmitted messages of the other user .
the users would like to communicate the open and secret messages with arbitrarily low probability of error , while maintaining secrecy of the secret messages .
the signals at the intended receiver and the wiretapper are given by [ eqn : tw ] @xmath41 where @xmath42 and @xmath29 .
we also assume the same power constraints given in ( with @xmath43 ) , and again use an equivalent standard form as illustrated in figure [ fig : tw ] : [ eqn : twstd ] @xmath44 where * the codewords @xmath45 are scaled to get @xmath46 and @xmath47 ; * the maximum powers are scaled to get @xmath48 and @xmath49 ; * the transmitters new channel gains are given by @xmath50 and @xmath51 ; * the wiretapper s new channel gains are given by @xmath52 and @xmath53 ; * the noises are normalized by @xmath54 and @xmath55 . in this section , we present some useful preliminary definitions including the secrecy constraint we will use . in particular , the secrecy constraint we used is the collective secrecy constraint " we defined in @xcite , and is suitable for the multi - access nature of the systems of interest .
we use the _ normalized joint conditional entropy _ of the transmitted messages given the eavesdropper s received signal as our secrecy constraint , i.e. , @xmath56 for any set @xmath57 of users . for perfect secrecy of all
transmitted secret messages , we would like @xmath58 assume @xmath59 for some arbitrarily small @xmath60 as required .
then , @xmath61 where @xmath62 as @xmath63 .
if @xmath64 , then we define @xmath65 .
thus , the perfect secrecy of the system implies the perfect secrecy of any group of users , guaranteeing that when the system is secure , so is each individual user .
[ def : achrate ] let @xmath66 .
the rate vector @xmath67 is said to be _ achievable _ if for any given @xmath68 there exists a code of sufficient length @xmath69such that [ eqn : achdef ] @xmath70 and @xmath71 is the average probability of error .
in addition , we need @xmath72 where @xmath73 denotes our secrecy constraint and is defined in .
we will call the set of all achievable rates , the _ secrecy - capacity region _ , and denote it @xmath74 for the ggmac - wt , and @xmath75 for the gtw - wt , respectively .
before we state our results , we also define the following notation which will be used extensively in the rest of this paper : @xmath76}}^+ } & \triangleq \max { { \left [ { \xi,0}\right ] } } \\ \label{eqn : capm } { c^{{{{\scriptscriptstyle \text{m}}}}}}_{{{\mathcal{s}}}}({{{\mathbf{p } } } } ) & \triangleq { \frac{1}{2}}\log { \left({1+{{\textstyle \sum}}_{k \in { { { \mathcal{s } } } } } p_k}\right ) } , \quad { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}\\ \label{eqn : capw } { c^{{{{\scriptscriptstyle \text{w}}}}}}_{{{\mathcal{s}}}}({{{\mathbf{p } } } } ) & \triangleq { \frac{1}{2}}\log { \left({1+{{\textstyle \sum}}_{k \in { { { \mathcal{s } } } } } h_k p_k}\right ) } , \quad { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}\\ \label{eqn : capws } { \tilde{c}^{{{{\scriptscriptstyle \text{w}}}}}}_{{{\mathcal{s}}}}({{{\mathbf{p } } } } ) & \triangleq { \frac{1}{2}}\log { \left({1+\frac{{{\textstyle \sum}}_{k \in { { { \mathcal{s } } } } } h_k p_k } { 1+{{\textstyle \sum}}_{k \in { { { { { \mathcal{s}}}}^c } } } h_k p_k}}\right ) } , \quad { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}\\ \label{eqn : pset } { { { \mathcal{p}}}}&\triangleq { { \left\ { { { { { \mathbf{p } } } } : 0 \le p_k \le { \bar{p}}_k , \ , \forall k}\right\}}}\\ \label{eqn : pmmax } { \bar{{{\mathbf{p}}}}}&\triangleq { { \left\ { { { \bar{p}}_1,\dotsc,{\bar{p}}_k}\right\}}}\end{aligned}\ ] ]
lastly , we informally call the @xmath26th user _ strong _ if @xmath77 , and _ weak _ if @xmath78 .
this is a way of indicating whether the intended receiver or the wiretapper is at a more of an advantage concerning that user , and is equivalent to stating whether the single - user secrecy capacity of that user is positive or zero .
we later extend this concept to refer to users who can achieve positive secrecy rates and those who can not .
in addition , we will say that a user is _ single - user decodable _ if its rate is such that it can be decoded by treating the other user as noise .
a user group @xmath79 is single - user decodable by the eavesdropper if @xmath80 .
our achievable rates can not guarantee secrecy for such a group of users .
in this section , we present our main results for the ggmac - wt . we first define two separate regions and then give an achievable region : let @xmath81 for all @xmath26 .
then , the superposition region , @xmath82 , is given by @xmath83}}^+ } , \ ; \forall { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}\bigr\ } \end{split}\ ] ] which can be written as @xmath84^+ \!\!,\ ; \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\biggr \}. \end{split}\ ] ] let @xmath85 be such that @xmath86 for all @xmath26 and @xmath87 .
let @xmath88 for all @xmath26 .
then , the tdma region , @xmath89 , is given by @xmath90}}^+ } \ !
, \ ; \forall k { \in } { { { \mathcal{k}}}}\bigr \ } \end{split}\ ] ] which is equivalent to @xmath91^+ \!\ ! , \forall k { \in } { { { \mathcal{k}}}}\biggr \}.\ ! \end{split } \ ] ] the superposition and tdma regions can also be written as follows : @xmath92}}^+ } , \quad \forall { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}\bigr \ } \end{split } \\ \begin{split } { { \ensuremath{{{\mathcal{g}}}^{\scriptscriptstyle \text{\text{ma - tdma}}}}\xspace}}({{{\mathbf{p}}}},{\boldsymbol{\alpha } } ) = \bigl \ { { { { \mathbf{r}}}}\colon \hspace{-1.2 in } & \\ & { { { r}^s}}_k + { { { r}^o}}_k \le \alpha_k { c^{{{{\scriptscriptstyle \text{m}}}}}}_k { \left({\frac{{\bar{p}}_k}{\alpha_k}}\right ) } , \quad \forall k \in { { { \mathcal{k}}}}\\ & { { { r}^s}}_k \le \alpha_k { { { \left [ { { c^{{{{\scriptscriptstyle \text{m}}}}}}_k { \left({\frac{{\bar{p}}_k}{\alpha_k}}\right ) } { - } { c^{{{{\scriptscriptstyle \text{w}}}}}}_k{\left({\frac{{\bar{p}}_k}{\alpha_k}}\right)}}\right ] } } ^+}\ ! , \ , \forall k \in { { { \mathcal{k}}}}\bigr \ } \end{split}\end{gathered}\ ] ] in accordance with the definitions in . [ thm : macach ]
the rate region given below is achievable for the ggmac - wt : @xmath93 we first show that the superposition encoding rate region given in for a fixed power allocation is achievable . consider the following coding scheme for rates @xmath94 for some @xmath95 : * superposition encoding scheme : * for each user @xmath26 , consider the following scheme : generate @xmath96 codebooks @xmath97 and @xmath98 .
@xmath99 consists of @xmath100 codewords , each component of which is drawn from @xmath101 .
codebook @xmath102 has @xmath103 codewords with each component randomly drawn from @xmath104 and @xmath98 has @xmath105 codewords with each component randomly drawn from @xmath106 where @xmath107 is an arbitrarily small number to ensure that the power constraints on the codewords are satisfied with high probability and @xmath108 .
define @xmath109 and @xmath110 . to transmit message @xmath111
, user @xmath26 finds the @xmath112 codewords corresponding to components of @xmath113 and also uniformly chooses a codeword @xmath114 from @xmath98 .
user @xmath26 then adds all these codewords and transmits the resulting codeword , @xmath115 , so that it actually transmits one of @xmath116 codewords .
let @xmath117 .
note that since all codewords are chosen uniformly , user @xmath26 essentially transmits one of @xmath118 codewords at random for each message @xmath5 , and its overall rate of transmission is @xmath119 .
specifically , we choose the rates to satisfy @xmath120^+ \ ! , \ , \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\!\end{aligned}\ ] ] which we can also write as : @xmath121^+\ ! , \quad \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}. \label{eqn : maccolsec}\end{aligned}\ ] ] note that if is zero for a group of users , we can not achieve secrecy for those users . when @xmath122 , if the sum - capacity of the main channel is less than that of the eavesdropper channel , i.e. , @xmath123 , secrecy is not possible for the system .
assume this quantity is positive . to ensure that we can mutually satisfy both ,
, we can reclassify some open messages as secret .
clearly , if we can guarantee secrecy for a larger set of messages , secrecy is achieved for the original messages . from the first set of conditions in and the gmac coding theorem , @xcite , with high probability the receiver can decode the codewords with low probability of error . to show the secrecy condition in , first note that
, the coding scheme described is equivalent to each user @xmath26 selecting one of @xmath100 messages , and sending a uniformly chosen codeword from among @xmath118 codewords for each .
define @xmath124 , and we have @xmath125 where we used @xmath126 , and thus we have @xmath127 to get .
we will consider the two terms individually .
first , we have the trivial bound due to channel capacity : @xmath128 now write @xmath129 since user @xmath26 independently sends one of @xmath118 codewords equally likely for each secret message , @xmath130 we can also write @xmath131 where @xmath132 as @xmath133 since , with high probability , the eavesdropper can decode @xmath134 given @xmath22 due to and code generation . using , , and in , we get @xmath135 now , let us consider the tdma region given in . this region is obtained when users who can achieve single - user secrecy use a single - user wire - tap code as in @xcite in a tdma schedule , where the time - share of each user @xmath26 is given by @xmath86 and @xmath87 . a transmitter @xmath26 who can achieve secrecy , i.e. , having @xmath136 , tranmits for @xmath137 portion of the time when all other users are silent , using @xmath138 power , satisfying its average power constraint over the tdma time - frame .
this approach was used in @xcite to achieve secrecy sum - capacity for individual constraints . when the channel is degraded , i.e. , @xmath139 for all @xmath140 , then for collective constraints the tdma region is seen to be a subset of the superposition region . however , this is not necessarily true for the general case , and by time - sharing between the two schemes we can generally achieve a larger achievable region , given in .
we remark that it is possible to further divide the open " messages to get more sets of private " messages which are also perfectly secret , i.e. , if we let @xmath141 , then as long as we impose the same restrictions on @xmath142 as @xmath143 , we can achieve perfect secrecy of @xmath144 , as in @xcite .
however , this does not mean that we have perfect secrecy at channel capacity , as the secrecy sub - codes carry information about each other . .
] observe that even for @xmath43 users , a rate point in this region is four dimensional , and hence can not be accurately drawn .
we can instead focus on the _ secrecy rate region _
, the region of all achievable @xmath143 .
the sub - regions @xmath145 are shown for different channel gains in figure [ fig : ggmacwtreg ] for fixed transmit powers , and @xmath43 users .
figure [ fig : ggmacwtureg ] represents how these regions change with different transmit powers when the channel gains are fixed . for the case
shown , we need the convex hull operation , as the achievable region is a combination of different superposition and tdma regions .
note also that the main extra condition for the superposition region is on the _ total extra randomness _ added . as a result , it is possible for stronger " users to help weak " users by contributing more to the necessary extra number of codewords , which is the sum - capacity of the eavesdropper .
such a weak user only has to make sure that it is not single - user decodable , provided the stronger users are willing to sacrifice some of their own rate and generate more superfluous codewords . in other words
, we see that users in a set @xmath79 are further protected from the eavesdropper by the fact that users in set @xmath146 are also undecodable , compared to the single - user case .
the tdma region , on the other hand , does not allow users to help each other this way .
as such , only users whose channel gains allow them to achieve secrecy on their own are allowed to transmit . .
] for the special degraded case of @xmath147 , the perfect secrecy rate region for @xmath148 becomes the region given by ( * ? ? ?
* theorem 1 ) for @xmath149 .
we also observe that even though there is a limit on the secrecy sum - rate achieved by our scheme , it is possible to send open messages to the intended receiver at rates such that the sum of the secrecy rate and open rate for all users is in the capacity region of the mac channel to the intended receiver . even though we can not send at capacity with secrecy , the codewords used to confuse the eavesdropper may be used to communicate meaningful information to the intended receiver . in this section ,
we present an achievable region for the gtw - wt using a superposition coding similar to that used to achieve the region @xmath82 for the ggmac - wt .
we first define let @xmath81 .
then , the gtw - wt superposition region , @xmath150 , is given by @xmath151^+ \!\ !
, \ , \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\biggr \ } \end{split}\ ] ] which can be written as @xmath152^+ \!\ ! , \ , \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\biggr \}. \ ! \end{split}\ ] ] we can also write this region more compactly as the following : @xmath153^+\ ! , \quad \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\biggr
\}. \end{split}\ ] ] [ thm : twach ] the rate region given below is achievable for the gtw - wt : @xmath154 . ]
the proof is very similar to the proof of theorem [ thm : macach ] .
we use the same coding scheme as theorem [ thm : macach ] , the main difference is that we choose the rates to satisfy @xmath155^+ \ ! , \ ; \forall { { { \mathcal{s}}}}{\subseteq } { { { \mathcal{k}}}}\end{aligned}\ ] ] or equivalently @xmath156^+ , \quad \forall { { { \mathcal{s}}}}\subseteq { { { \mathcal{k}}}}.\end{aligned}\ ] ] assuming is positive .
the decodability of @xmath22 from @xmath157 comes from and the capacity region of the gaussian two - way channel @xcite .
this gives the first set of terms in the achievable region .
the key here is that since each transmitter knows its own codeword , it can _ subtract its self - interference _ from the received signal and get a clear channel .
therefore , the gaussian two - way channel decomposes into two parallel channels .
the second group of terms in , resulting from the secrecy constraint , can be shown the same way as the proof of theorem [ thm : macach ] , since @xmath158 has the same form for both channels .
in other words , as far as the eavesdropper is concerned , the channel is still a gmac with @xmath43 users . as such
, we need to send @xmath159 extra codewords in total , which need to be shared by the two - terminals provided they are not single - user decodable . .
] for different channel gains , the region of all @xmath143 satisfying is shown in figure [ fig : gtwwtreg ] .
since we require four dimensions for an accurate depiction of the complete rate region , we only focus on our main interest , i.e. , the secrecy rate region .
figure [ fig : gtwwtureg ] shows the achievable secrecy rate region as a function of transmit powers .
we note that higher powers always result in a larger region .
we indicate the constraint on the overall rates , corresponding to the capacity region of the gaussian two - way channel , by the dotted line .
note that the secrecy region has a structure similar to the ggmac - wt with @xmath43 .
as far as the eavesdropper is concerned , there is no difference between the two channels .
however , since the main channel between users decomposes into two parallel channels , higher rates can be achieved between the legitimate terminals ( users ) .
thus , in effect , each user s transmitted codewords act as a _ secret key _ for the other user s transmitted codewords , requiring fewer extraneous codewords overall to confuse the eavesdropper , and a larger secrecy region .
we note that a user may either achieve secrecy or not , depending on whether it is single - user decodable or not . as a result
, tdma does not enlarge the region , since each user can at least achieve their single - user secrecy rates . to see this , note that the constraint on the secrecy sum - rate can be written as : @xmath160 so that transmitting in the two - way channel always provides an advantage over the single - user channels .
the achievable regions given in theorems [ thm : macach ] and [ thm : twach ] depend on the transmit powers .
we are , thus , naturally interested in the power allocation @xmath161 that would maximize the total secrecy sum - rate .
recall that the standardized channel gain for user @xmath26 is @xmath162 , and that the higher @xmath163 is , the better the corresponding eavesdropper channel . without loss of generality , assume that users are ordered in terms of increasing standardized eavesdropper channel gains , i.e. , @xmath164 .
note that , we only need to concern ourselves with the case @xmath165 , since we can combine users with the same channel gains into one super - user .
we can then split the resulting optimum power allocation for a super - user among the actual constituting users in any way we choose , since they would all result in the same sum - rate .
in addition , from a physical point of view , assuming that the channel parameters are drawn according to a continuous distribution and then fixed , the probability that two users would have the same exact standardized channel gain is zero .
we first examine the superposition region given in .
the secrecy sum - rate achievable with superposition coding for the ggmac - wt was given in theorem [ thm : macach ] as @xmath166}}^+ } \!\!\ ] ] and we would like to find the power allocation that maximizes this quantity .
stated formally , we are interested in the transmit powers that solve the following optimization problem : @xmath167 } } \notag \hspace{-1.5 in } & \\ & = \min_{{{{\mathbf{p}}}}\in { { { \mathcal{p } } } } } \ ; { \frac{1}{2}}\log { \phi}_{{{\mathcal{k}}}}({{{\mathbf{p } } } } ) \\ & \equiv \min_{{{{\mathbf{p}}}}\in { { { \mathcal{p } } } } } \ ; { \phi}_{{{\mathcal{k}}}}({{{\mathbf{p } } } } ) \label{eqn : macsumprob1}\end{aligned}\ ] ] where @xmath168 and @xmath122 yields . in obtaining , we simply used the monotonicity of the @xmath169 function .
the solution to this problem is given below : [ thm : summac ] the secrecy sum - rate maximizing power allocation for @xmath82 satisfies @xmath170 if @xmath171 and @xmath172 is @xmath173 where @xmath174 is some limiting user satisfying @xmath175 and we define @xmath176 . note that this allocation shows that only a _ subset of the strong users _ must be transmitting .
we start with writing the lagrangian to be minimized : @xmath177 equating the derivative of the lagrangian to zero , we get @xmath178 where we define @xmath179 for any set @xmath57 . it is easy to see that if @xmath180 , then @xmath181 , and we have @xmath182 . if @xmath183 , then we similarly find that @xmath184 . finally , if @xmath185 , then we also have @xmath186 and @xmath187 does not depend on @xmath188 , so we can set @xmath184 with no effect on the secrecy sum - rate . thus , we have @xmath182 if @xmath183 , and @xmath184 if @xmath189 .
then , the optimal set of transmitters is of the form @xmath190 since if a user @xmath191 is transmitting , all users such that @xmath192 must also be transmitting .
we also note that @xmath193 .
let @xmath191 be the last user satisfying this property , i.e. @xmath194 and @xmath195 .
note that @xmath196 in other words , all sets @xmath197 for @xmath198 also satisfy this property , and are viable candidates for the optimal set of transmitting users .
therefore , we can claim that @xmath199 is the optimum set of transmitting users , since from above we can iteratively see that @xmath200 for all @xmath201 . note that , for the special case of @xmath43 users , the optimum power allocation is @xmath202 we also need to consider the tdma region . in this case , the maximum achievable secrecy sum - rate is : @xmath203}}. \ ] ] this is a simple complex optimization problem that can easily be solved numerically .
for the degraded case , we can obtain a closed form solution : @xmath204 as in @xcite .
in general , we can not obtain such a solution .
however , it is trivial to note that users with @xmath205 should not be transmitting in this scheme .
the secrecy sum - rate is then the maximum of the solutions given by the superposition and tdma regions .
now , we will examine the power allocation that maximizes the secrecy sum - rate given in theorem [ thm : twach ] as @xmath206^+ \!.\ ] ] this problem is formally stated below : @xmath207 } } \\
\equiv \min_{{{{\mathbf{p}}}}\in { { { \mathcal{p } } } } } \ ; { \psi}_{{{\mathcal{k}}}}({{{\mathbf{p } } } } ) \label{eqn : twsumprob1}\end{gathered}\ ] ] where @xmath208 and @xmath122 yields .
the optimum power allocation is stated below : [ thm : sumtw ] the secrecy sum - rate maximizing power allocation for the gtw - wt is given by @xmath209 the lagrangian is , @xmath210 equating the derivative of the lagrangian to zero for user @xmath211 , we get @xmath212 where @xmath213 an argument similar to the one for the ggmac - wt establishes that if @xmath214 , or equivalently if @xmath215 , then @xmath216 .
when equality is satisfied , then @xmath217 regardless of @xmath188 , and as such @xmath218 can be seen to not depend on @xmath188 .
to conserve power , we again set @xmath219 in this case . on the other hand , if @xmath220 , then @xmath182 .
consider user 1 .
if @xmath221 , and @xmath222 , this implies that @xmath223 . since @xmath224
, we can not have @xmath221 . as a consequence of this contradiction
, we see that @xmath225 whenever @xmath221 .
assume @xmath226 , and consider the two alternatives for @xmath227 .
we will have @xmath225 if @xmath228 ; and @xmath229 if @xmath230 .
these cases correspond to @xmath231 and @xmath232 , respectively .
thus , we have as the secrecy sum - rate maximizing power allocation .
observe that the solution in theorem [ thm : sumtw ] has a structure similar to that in theorem [ thm : summac ] . in summary
, it is seen that as long as a user is not single - user decodable , it should be transmitting with maximum power .
hence , when both users can be made to be non - single - user decodable , then the maximum powers will provide the largest secrecy sum - rate .
if this is not the case , then the user who is single - user decodable can not transmit with non - zero secrecy and will just make the secrecy sum - rate constraint tighter for the remaining user by transmitting open messages . comparing to , we see that the same form of solutions is found , but the range of channel gains where transmission is possible is larger , showing that gtw - wt allows secrecy even when the eavesdropper s channel is not very weak .
in the previous section , we found the secrecy sum - rate maximizing power allocations .
for both the ggmac - wt and gtw - wt , if the eavesdropper is not disadvantaged enough " for some users ,
then these users transmit powers are set to zero .
we posit that such a user may be able to help " a transmitting user , since it can cause more harm to the eavesdropper than to the intended receiver .
we only consider the superposition region , since in the tdma region a user has a dedicated time - slot , and hence does not affect the others .
we will next show that this type of cooperative behavior is indeed useful , notably exploiting the fact that the established achievable secrecy sum - rate is a difference of the sum - capacity expressions for the intended channel(s ) and the eavesdropper s channel . as a result , reducing the latter more than the former actually results in an _ increase _ in the achievable secrecy sum - rate .
formally , the scheme we are considering implies partitioning the set of users , @xmath233 into a set of transmitting users , @xmath199 and a set of jamming users @xmath234 . if a user @xmath26 is jamming , then it transmits @xmath235 instead of codewords . in this case
, we can show that we can achieve higher secrecy rates when the weaker " users are jamming .
we also show that the gtw - wt , has an additional advantage compared to the ggmac - wt , that is the fact that the receiver already knows the jamming sequence . as such
, this scheme only harms the eavesdropper and not the intended receivers , achieving an even higher secrecy sum - rate . once again , without loss of generality , we consider @xmath165 .
in addition , we will assume that a user can either take the action of transmitting its information or jamming the eavesdropper , but not both .
it is readily shown in section [ sec : jammac ] below that we do not lose any generality by doing so , and that splitting the power of a user between the two actions is suboptimal from the secrecy sum - rate maximization point of view . the problem is formally presented below : @xmath236\\ & \hspace{.2 in } \equiv \label{eqn : jammac } \min_{{{{\mathcal{t}}}}\subseteq { { { \mathcal{k } } } } , { { { \mathbf{p}}}}\in { { { \mathcal{p } } } } } \frac{{\phi}_{{{\mathcal{k}}}}({{{\mathbf{p}}}})}{{\phi}_{{{{{\mathcal{t}}}}^c}}({{{\mathbf{p}}}})}\end{aligned}\ ] ] where we recall that @xmath237 is given by , such that @xmath238 to see that a user should not be splitting its power among jamming and transmitting , it is sufficient to note that regardless of how a user splits its power , @xmath239 will be the same , and the user only affects @xmath240 .
assume the optimum solution is such that user @xmath211 splits its power , so @xmath241 and @xmath242 .
then , it is easy to see that if @xmath243 , the sum - rate is increased when that user uses its jamming power to transmit , and when @xmath244 , the sum - rate is increased when the user uses its transmit power to jam . when @xmath245 , then regardless of how its power is split , the sum - rate is the same , and we can assume user @xmath211 either transmits or jams .
note that we must have @xmath246 to have a non - zero secrecy sum - rate , and @xmath247 to have an advantage over not jamming .
this scheme can be shown to achieve the following secrecy sum - rate : [ thm : jammac ] the secrecy sum - rate using cooperative jamming is @xmath248 where @xmath199 is the set of transmitters and the optimum power allocation is of the form @xmath249 with @xmath250}}^+}\ ] ] and @xmath251 whenever the positive real root exists , and @xmath252 otherwise .
we first solve the subproblem of finding the optimal power allocation for a set of given transmitters , @xmath199 .
the solution to this will also give us insight into the structure of the optimal set of transmitters , @xmath253 .
we start with writing the lagrangian : @xmath254 the derivative of the lagrangian depends on the user : @xmath255 since a user @xmath242 satisfies @xmath256 , it must have @xmath257 .
consider a user @xmath241 .
the same argument as in the sum - rate maximization proof leads to @xmath182 if @xmath183 and @xmath221 if @xmath189 .
now examine a user @xmath242 .
we can write as @xmath258 where @xmath259 let @xmath260 then , we have @xmath261 iff @xmath262 , and @xmath263 iff @xmath264 .
thus , we again find that we must have @xmath265 for all @xmath242 .
also , if @xmath266 , then @xmath182 .
only if @xmath267 , can we have @xmath268 .
now , since @xmath269 , we must have @xmath270 .
thus , we find that @xmath271 .
then , we know that for a given set of transmitters , @xmath199 , the solution is such that all users @xmath241 transmit with power @xmath272 if @xmath273 . in the set of jammers @xmath274 , all users have @xmath275 , and when this inequality is not satisfied with equality , the jammers jam with maximum power .
if the equality is satisfied for some users @xmath211 , their jamming powers can be found from solving @xmath276 . by rearranging terms in ,
we note that the optimum power allocation for this user , call it user @xmath277 , is found by solving the quadratic @xmath278 the solution of which is given in .
note that defines an ( upright ) parabola .
if the root given in exists and is positive , then @xmath279 .
this comes from the fact that if @xmath280 , then @xmath281 for all @xmath282 , and we must have @xmath283 .
if , on the other hand , gives a complex or negative solution , then the parabola does not intersect the @xmath284 axis , and is always positive .
hence , @xmath285 , and @xmath277 does not belong to @xmath274 , i.e. @xmath286 . the form of this solution is intuitively pleasing , since it makes more sense for weaker " users to jam as they harm the eavesdropper more than they do the intended receiver .
what we see is that all transmitting users @xmath211 , such that @xmath256 , transmit with maximum power as long as their standardized channel gain @xmath287 is less than some limit @xmath288 , and all jamming users must have @xmath180 .
we claim that all users in @xmath253 must have @xmath289 and all users in @xmath290 have @xmath265 . to make this argument
, we need to show that a @xmath199 such that there exists some @xmath291 with @xmath292 and @xmath293 such that @xmath294 can not be the optimum set . to see this ,
let @xmath161 be the optimum power allocation for a set @xmath199 .
consider a new power allocation and set such that @xmath295 , i.e. , user @xmath296 is now jamming , and let @xmath297 , @xmath298 and @xmath299 , for some small @xmath300 .
we then have @xmath301 which is a lower value for the objective function , proving that @xmath302 is not optimum .
this shows that all users @xmath303 must have @xmath304 for all users @xmath305 .
since the last user in @xmath290 has @xmath276 , necessarily @xmath265 for all @xmath303 , and @xmath289 for all @xmath306 . summarizing , the optimum power allocation is such that there is a set of transmitting users @xmath307 with @xmath170 for @xmath308 , there is a set of silent users @xmath309 , and a set of jamming users @xmath310 with @xmath170 for @xmath311 and @xmath312 is found from @xmath313 .
this is what is presented in the statement in theorem [ thm : jammac ] .
note that to find @xmath314 , we can simply do an exhaustive search as we have narrowed the number of possible optimal sets to @xmath315 instead of @xmath316 and found the optimal power allocations for each .
* two - user ggmac - wt : * for illustration purposes , let us consider the familiar case with @xmath43 transmitters . in this case , we know that either user 2 jams , or no user does . the solution can be found from comparing the two cases .
if , without jamming , user @xmath112 can transmit , then it is optimal for it to continue to transmit , and jamming will not improve the sum - rate .
otherwise , user 2 may be jamming to improve the secrecy rate of user 1 .
the optimum power allocation for user 1 is equivalent to @xmath317 if @xmath318 and @xmath221 if @xmath319 .
the power for user @xmath112 is found from .
for @xmath112 users , we can simply write as @xmath320 where @xmath321}}.\end{aligned}\ ] ] with different @xmath322 for @xmath323 .
the circles indicate optimum jamming power . ] if @xmath231 , we automatically have @xmath317 . in addition , we have @xmath324 , so we only need to concern ourselves with the possibly positive root , @xmath325 .
we first find when @xmath225 .
we see that @xmath326 for all @xmath95 if @xmath327 , equivalent to having two negative roots , or @xmath328 , equivalent to having no real roots of @xmath329 .
now examine when @xmath330 .
this is possible if and only if @xmath331 .
since @xmath332 , this happens only when @xmath333 or @xmath334 .
however , if @xmath333 , we are better off transmitting than jamming .
the last case to examine is when @xmath335 .
this implies that @xmath336 , and is satisfied when @xmath337 .
assume @xmath338 . in this case ,
we are guaranteed @xmath339 .
if @xmath221 , then we must have @xmath225 since the secrecy rate is 0 .
we would like to find when we can have @xmath332 . since @xmath340 , we must have @xmath341 , and @xmath342 .
this implies @xmath343 .
it is easy to see that @xmath344 if @xmath345 and @xmath225 otherwise
. thus , for @xmath43 users , the solution simplifies to : @xmath346}}^+ } ) , & \text{if } h_1 { \le } 1 , h_2 { > } 1\\ ( 0,0 ) , & \text{if } h_1 { \ge } 1 , h_2 { \ge } h_1 , \frac{h_1 - 1}{h_2-h_1}{\ge } { \bar{p}}_1 \\ ( { \bar{p}}_1,\min { { \left\ { { p,{\bar{p}}_2}\right\ } } } ) , & \text{if } h_1 { \ge } 1 , h_2 { > } h_1 , \frac{h_1 - 1}{h_2-h_1 } { < } { \bar{p}}_1 \end{cases } \hspace{-\multlinegap}\end{gathered}\ ] ] where @xmath347}}}}{h_2 ( h_2{-}h_1)}.\ ] ] this solution can be checked to be in accordance with the sum - rate maximizing power allocation of theorem [ thm : summac ] .
we note that in the case unaccounted for in , i.e. , when @xmath348 and @xmath349 , both users should be transmitting . in general
, the solution shows that the weaker " user should jam if it is not single - user decodable , and if it has enough power to make the other user strong " in the new effective channel . with different @xmath322 for @xmath350 .
the circles indicate optimum jamming power . ] once again , we propose to maximize the secrecy sum - rate using cooperative jamming when useful .
this problem is formally stated as follows : @xmath351}}\\ \equiv \min_{{{{\mathcal{t}}}}\subseteq { { { \mathcal{k } } } } } \min_{{{{\mathbf{p}}}}\in { \bar { { { \mathcal{p } } } } } } \frac{{\psi}_{{{\mathcal{k}}}}({{{\mathbf{p}}}})}{{\psi}_{{{{{\mathcal{t}}}}^c}}({{{\mathbf{p } } } } ) } \label{eqn : jamtw}\end{gathered}\ ] ] where we recall that @xmath352 is given by and @xmath353 note that @xmath354 since there are only two terminals . a similar argument to the ggmac - wt case
can easily be used to establish that we can assume a user to be either transmitting or jamming , but not both . since the jamming user is also the receiver that the other user is communicating with and knows the transmitted signal , this scheme entails no loss of capacity as far as the transmitting user is concerned .
the optimum power allocations are given as follows .
[ thm : jamtw ] the achievable secrecy sum - rate for the the collaborative scheme described is @xmath355 where @xmath199 is the set of transmitting users and the optimum power allocations are given by @xmath356 & 1{<}h_1{<}1{+}h_2{\bar{p}}_2 , \\ & { \psi}_2({\bar{{{\mathbf{p}}}}}){>}{\psi}_1({\bar{{{\mathbf{p } } } } } ) \end{aligned } \\ \ !
( { \bar{p}}_1,{\bar{p}}_2 ) , \ , \text{2 transm . , 1 jams , } & \!\!\text{if } \begin{aligned}[t ] & 1{<}h_1{<}h_2{<}1{+}h_1{\bar{p}}_1 , \\ & { \psi}_1({\bar{{{\mathbf{p}}}}}){>}{\psi}_2({\bar{{{\mathbf{p } } } } } ) \end{aligned}\\ \ ! ( 0,0 ) , & \!\ !
\text{otherwise } \end{cases } \hspace{-\multlinegap}\end{gathered}\ ] ] similar to the ggmac - wt , we start with the sub - problem of finding the optimal power allocation given a jamming set .
the lagrangian is given by @xmath357 taking the derivative we have @xmath358 since a user @xmath242 satisfies @xmath256 , it must have @xmath257 .
consider user @xmath241 .
we again argue that if @xmath359 , then @xmath184 and if @xmath360 , then @xmath182 .
now examine a user @xmath242 .
it is easy to see that since such a user only harms the jammer , the optimal jamming strategy should have @xmath182 , i.e. , the maximum power .
this can also be seen by noting that for this case simplifies to @xmath361 and hence we must have @xmath362 for all @xmath242 . the jamming set will be one of @xmath363 , since there is no point in jamming when there is no transmission . also , if any of the two users is jamming , by the argument above , @xmath182 , @xmath364 .
we can easily see that jamming by a user @xmath211 only offers an advantage if @xmath365 , i.e. , @xmath366 iff @xmath365 for @xmath242 .
thus , when @xmath367 , both users should be transmitting instead of jamming . however , when any user has @xmath365 , jamming always does better than the case when both users are transmitting . in this case ,
@xmath368 for some user @xmath211 , and the objective function in is minimized when this user is jamming , and the other one is transmitting .
if , however , @xmath369 , then it will not transmit , and we should not be jamming . consolidating all of these results ,
we come up with the power allocation in in theorem [ thm : jamtw ] . with different @xmath322 for @xmath370 .
] a sufficient , but not necessary condition for the weaker user to be the jamming user is if @xmath371 ; this case corresponds to having higher snr at the eavesdropper for the original , non - standardized model .
this can be interpreted as jam with maximum power if it is possible to change user 1 s effective channel gain such that it is no longer single - user decodable " . for the simple case of equal power constraints , @xmath372
, it is easily seen that user 1 should never be jamming .
the optimal power allocation in that case reduces to @xmath373
in this section , we present numerical results to illustrate the achievable rates obtained , as well as the cooperative jamming scheme and its effect on achievable secrecy sum - rates .
as mentioned earlier in the paper , examples of achievable secrecy rate regions are given in figures [ fig : ggmacwtureg ] and [ fig : gtwwtureg ] for the ggmac - wt with @xmath43 and gtw - wt respectively .
comparing figures [ fig : ggmacwtureg ] and [ fig : gtwwtureg ] , we see that the gtw - wt achieves a larger secrecy rate region then the ggmac - wt , and offers more protection to weak " users . in addition , tdma does not enlarge the achievable region for gtw - wt since superposition coding always allows users to achieve their single - user secrecy rates for any transmit power .
let us have a closer look at the secrecy advantage of the two - way channel over the mac with two users .
for the ggmac - wt with @xmath43 , the achievable maximum secrecy sum - rate , @xmath374 is limited by the channel parameters .
it was shown in @xcite that for the degraded case , @xmath375 , the secrecy sum - capacity , @xmath376 , is an increasing function of the total sum power , @xmath377 .
however , it is limited since @xmath378 as @xmath379 . for the general case , where @xmath380 , theorem [ thm : summac ] implies that the sum - rate is maximized when only user 1 transmits ( assuming @xmath381 ) , and is bounded similarly by @xmath382 . on the other hand , for the gtw - wt , unlike the ggmac - wt , it is possible to increase the secrecy capacity by increasing the transmit powers .
this mainly stems from the fact that the users now have the extra advantage over the eavesdropper that they know their own transmitted codewords . in effect
, _ each user helps encrypt the other user s transmission_. to see this more clearly , consider the symmetric case where @xmath383 and @xmath384 , which makes all users receive a similarly noisy version of the same sum - message .
the only disadvantage the eavesdropper has , is that he does not know any of the codewords whereas user @xmath26 knows @xmath115 . in this case
, @xmath385 is achievable , and this rate approaches @xmath386 as @xmath387 .
thus , it is possible to achieve a secrecy - rate increase at the same rate as the increase in channel capacity .
next , we examine the secrecy sum - rate maximizing power allocations and optimum powers for the cooperative jamming scheme .
figures [ fig : ggmacwtjamp2 ] and [ fig : ggmacwtjamp100 ] show the achievable secrecy rate improvement for the cooperative jamming scheme for various channel parameters for the ggmac - wt with @xmath43 .
the plots are the secrecy rates for user 1 when user 2 is jamming with a given power , which correspond to user 1 s single - user secrecy _ capacity _ , @xcite , since only one user is transmitting .
when @xmath388 , the secrecy capacity is seen to be zero , unless user @xmath112 has enough power to _ convert _ user 1 s re - standardized channel gain to less than 1 . for the gtw - wt , it is always optimal for user @xmath112 to jam as long as it enables user @xmath389 to transmit , as seen in figure [ fig : twwtjamp2 ] .
the results show , as expected , that secrecy is achievable for both users so long as we can keep the eavesdropper from single - user decoding the transmitted codewords by treating the remaining user as noise .
since the coding schemes considered here assume knowledge of eavesdropper s channel gains , applications are limited .
one practical application could be securing of a physically protected area such as inside a building , when the eavesdropper is known to be outside .
in such a case we can design for the worst case scenario .
an example is given in figure [ fig : ggmacwtcoljam ] for the ggmac - wt , where we assume a simple path - loss model and fixed locations for two transmitters ( t ) and one receiver ( r ) at the center .
we examine the transmit / jam powers for this area when the eavesdropper is known to be at @xmath390 using a fixed path - loss model for the channel gains , and plot the transmit / jam powers and the achieved secrecy sum - rates as a function of the eavesdropper location .
it is readily seen that when the eavesdropper is close to the bs , the secrecy sum - rate falls to zero . also , when the eavesdropper is in the vicinity of a transmitter , that transmitter can not transmit in secrecy .
however , in this case , the transmitter can jam the eavesdropper effectively , and allow the other transmitter to transmit and/or increase its secrecy rate with little jamming power .
the situation for the gtw - wt is similar , and is shown in figure [ fig : gtwwtcoljam ] . in this case , jamming is more useful as compared to the ggmac - wt , and we see that it is possible to provide secrecy for a much larger area where the eavesdropper is located , as the jamming signal does not hurt the intended receiver .
in this paper , we have considered the gaussian multiple access and two - way channels in the presence of an external eavesdropper who receives the transmitted signals through a multiple - access channel , and provided achievable secrecy rates .
we have shown that the multiple - access nature of the channels considered can be utilized to improve the secrecy of the system .
in particular , we have shown that the total extra randomness is what matters mainly concerning the eavesdropper , rather than the individual randomness in the codes .
as such , it may possible for users whose single - user wire - tap capacity are zero , to communicate with non - zero secrecy rate , _ as long as it is possible to put the eavesdropper at an overall disadvantage_. this is even clearer for two - way channels , where even though the eavesdropper s channel gain may be better than a terminal s , the extra knowledge of its own codeword by that terminal enables communication in perfect secrecy as long as the eavesdropper s received signal is not strong enough to allow single - user decoding .
we found achievable secrecy rate regions for the general gaussian multiple - access wire - tap channel ( ggmac - wt ) and the gaussian two - way wire - tap channel ( gtw - wt ) .
we also showed that for the ggmac - wt the secrecy sum - rate is maximized when only users with strong " channels to the intended receiver as opposed to the eavesdropper transmit , and they do so using all their available power .
for the gtw - wt , the sum - rate is maximized when both terminals transmit with maximum power as long as the eavesdropper s channel is not good enough to decode them using single - user decoding . finally , we proposed a scheme termed _ cooperative jamming _ , where a disadvantaged user may help improve the secrecy rate by jamming the eavesdropper .
we found the optimum power allocations for the transmitting and jamming users , and showed that significant rate gains may be achieved , especially when the eavesdropper has much higher snr than the receivers and normal secret communications is not possible .
the gains can be significant for both the ggmac - wt and gtw - wt .
this cooperative behavior is useful when the maximum secrecy sum - rate is of interest .
we have also contrasted the secrecy rates of the two channels we considered , noting the benefit of the two - way channels where the fact that each receiver has perfect knowledge of its transmitted signal brings an advantage with each user effectively encrypting the communications of the other user . in this paper , we only presented achievable secrecy rates for the ggmac - wt and gtw - wt .
the secrecy capacity region for these channels are still open problems . in @xcite
, we also found an upper bound for the secrecy sum - rate of the ggmac - wt and noted that the achievable secrecy sum - rate and the upper bound we found only coincide for the degraded case , so that we have the secrecy sum - capacity for the degraded gmac - wt . even though there is a gap between the achievable secrecy sum - rates and upper bounds , cooperative jamming
was shown in @xcite to give a secrecy sum - rate that is close to the upper bound in general .
finally , we note that the results provided are of mainly theoretical interest , since as of yet there are no currently known practical codes for multi - access wire - tap channels unlike the single - user case where in some cases practical codes have been shown to be useful for the wire - tap channel , @xcite .
furthermore , accurate estimates of the eavesdropper channel parameters are required for code design for wire - tap channels where the channel model is quasi - static , as in our models considered in this paper .
e. tekin and a. yener , `` achievable rates for the general gaussian multiple access wire - tap channel with collective secrecy , '' in _ proc .
allerton conf .
_ , monticello , il , sep 2729 2006 , [ online . ]
available : http://arxiv.org/abs/cs/0612088 .
u. maurer and s. wolf , `` information - theoretic key agreement : from weak to strong secrecy for free , '' in _ proceedings of eurocrypt 2000 , lecture notes in computer science _ ,
1807.1em plus 0.5em minus 0.4em springer - verlag , 2000 , pp .
351368 .
u. maurer and s. wolf , `` secret - key agreement over unauthenticated public channels - part i : definitions and a completeness result , '' _ ieee trans .
inform . theory _
49 , no . 4 , pp . 822831 , apr 2003 .
r. liu , i. maric , r. d. yates , and p. spasojevic , `` discrete memoryless interference and broadcast channels with confidential messages , '' in _ proc .
allerton conf .
_ , monticello , il , sep 27 - 29 2006 .
y. liang and v. poor , `` generalized multiple access channels with confidential messages , '' _ ieee trans .
inform . theory _ , submitted for publication , [ online .
] available : http://arxiv.org/format/cs.it/0605014 .
r. liu , i. maric , r. d. yates , and p. spasojevic , `` the discrete memoryless multiple access channel with confidential messages , '' in _ proc .
inform . theory ( isit )
_ , seattle , wa , jul 9 - 14 2006 .
a. thangaraj , s. dihidar , a. r. calderbank , s. mclaughlin , and j .-
merolla , `` applications of ldpc codes to the wiretap channel , '' _ ieee trans .
53 , no .
29332945 , aug 2007 .
m. bloch , j. barros , m. r. d. rodrigues , and s. w. mclaughlin , `` wireless information - theoretic security - part ii : practical implementation , '' _ ieee trans .
inform . theory _ ,
submitted for publication . | the general gaussian multiple access wire - tap channel ( ggmac - wt ) and the gaussian two - way wire - tap channel ( gtw - wt ) are considered . in the ggmac - wt ,
multiple users communicate with an intended receiver in the presence of an eavesdropper who receives their signals through another gmac . in the gtw - wt ,
two users communicate with each other over a common gaussian channel , with an eavesdropper listening through a gmac . a secrecy measure that is suitable for this multi - terminal environment
is defined , and achievable secrecy rate regions are found for both channels . for both cases , the power allocations maximizing the achievable secrecy sum - rate are determined .
it is seen that the optimum policy may prevent some terminals from transmission in order to preserve the secrecy of the system .
inspired by this construct , a new scheme , _ cooperative jamming _
, is proposed , where users who are prevented from transmitting according to the secrecy sum - rate maximizing power allocation policy jam " the eavesdropper , thereby helping the remaining users .
this scheme is shown to increase the achievable secrecy sum - rate .
overall , our results show that in multiple - access scenarios , users can help each other to collectively achieve positive secrecy rates .
in other words , cooperation among users can be invaluable for achieving secrecy for the system .
secrecy capacity , gaussian multiple access channel , gaussian two - way channel , wire - tap channel , confidential messages |
first introduced in 1954 @xcite , the notion of a positive basis plays a key role in understanding derivative free optimization methods that use a direct search framework ( a.k.a .
pattern search methods ) @xcite . in loose terms , a positive basis is a set that provides some directional information into every half - space , without including unnecessary vectors .
pattern search methods seek a minimizer of @xmath5 by evaluating @xmath5 over an ever shrinking ` pattern ' : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given @xmath5 and a incumbent solution @xmath6 .
evaluate @xmath7 for each @xmath8 .
+ if a lower function value is found , then update @xmath6 .
+ if no lower function value is found , then shrink @xmath9 . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if @xmath10 , @xmath11 , and the pattern forms a positive basis , then at least one vector in the pattern is a descent direction for @xmath5 at @xmath6 .
thus , under these conditions , once @xmath9 is sufficiently small , descent must occur .
formalization of this idea has been used to prove convergence in a number of direct search / pattern search methods @xcite ( and many others ) .
mathematically , to define positive basis we begin with the positive span : given a set of vectors @xmath12 , we define the positive span of @xmath13 by @xmath14 we now define @xmath13 to be a positive basis if it satisfies two properties 1 .
@xmath15 , and 2 .
@xmath16 for all @xmath17 . unlike bases of @xmath0 ( which always have exactly @xmath18 elements )
, a positive basis of @xmath0 has between @xmath1 and @xmath2 elements .
the lower bound of @xmath1 comes from the following fact .
* thm 3.7 ) if @xmath19 is a positive spanning set of @xmath0 , then @xmath20 is a spanning set of @xmath0 .
thus , to satisfy positive spanning condition , a set must contain at least @xmath1 vectors .
the upper bound on the cardinality of a positive basis is more subtle to obtain .
the original work of davis includes a proof of this fact ( * ? ? ?
* thm 6.7 ) , but the derivation is rather technical .
alternate proofs appear in the works of shephard @xcite and audet @xcite .
it is important to note that all three proofs use _ both _ positive spanning and positive linear independence to provide the upper bound . as such
, the property of positive linear independence is less clear in its implications on cardinality .
for example , any set consisting of exactly one vector is always positive linearly independent including the set @xmath21 .
( of course , any positive linearly independent set with @xmath22 or more vectors can not contain the @xmath23 vector . )
based on a knowledge of positive bases , one might conjecture that positively linearly independent sets have at most @xmath2 elements .
this is true in @xmath24 and @xmath25 , but false beyond that . in this note , we provide details proving that a positively linearly independent set in @xmath0 for @xmath3 has at most @xmath2 elements , but a positively linearly independent set in @xmath0 for @xmath4 can exceed this bound .
proofs of the first two statements appear in section [ lowdim ] . in section [ highdim ]
we provide two examples .
the first is a straightforward example of a positively linearly independent set @xmath13 with @xmath26 this example requires no tools other than basic linear algebra .
the second example develops a positively linearly independent set in @xmath0 with an arbitrary number of elements .
this example requires a small knowledge of convex analysis , but is still quite accessible to a general audience .
we conclude the introduction with a lemma that allows us to simplify notation in all future proofs and examples .
[ lem : normalize ] let @xmath27 with @xmath28 for all @xmath29 .
then @xmath13 is positively linearly independent if and only if @xmath30 is positively linearly independent .
note that @xmath31 where @xmath32 and @xmath33 for any @xmath34 .
thus , @xmath35 is positively linearly independent if and only if @xmath36 is positively linearly independent .
the maximal cardinality of a positively linearly independent set in @xmath24 is 2 .
suppose @xmath37 .
if @xmath38 , then @xmath13 is not positively linearly independent . if @xmath39 , then , by lemma [ lem : normalize ] , without loss of generality @xmath40 for each @xmath29 . by the pigeon hole principle , at least two vectors @xmath29 and @xmath41 are the same , and @xmath13 is not positively linearly independent . thus , if @xmath13 is positively linearly independent , then @xmath42 .
[ thm : r2]the maximal cardinality of positively linearly independent sets in @xmath25 is 4 . for eventual contradiction ,
suppose @xmath43 is positively linearly independent .
without loss of generality , assume @xmath44 for @xmath45 . clearly , @xmath46 for any @xmath47 . moreover , if @xmath48 for some @xmath29 , then @xmath49 for any @xmath50 .
as such , there are at least two vectors , say @xmath51 and @xmath52 , that are linearly independent . examining , @xmath53 , and repeating the same arguments
, we may assume that @xmath54 and @xmath55 are also linearly independent .
therefore , @xmath56 and @xmath57 are both bases of @xmath25 . as @xmath56 is a basis of @xmath25 , there exists @xmath58 such that @xmath59 .
if @xmath60 and @xmath61 , then we have contradicted the positive linear independence of @xmath13 . if @xmath62 and @xmath63 , then @xmath64 , which also contradicts positive linear independence . similarly , @xmath65 and @xmath66 fails
hence , @xmath67 . using the same arguments , and
@xmath57 is a basis of @xmath25 , there exists @xmath68 and @xmath69 such that @xmath70 and @xmath71 substitution now yields @xmath72 which shows that @xmath73 , so @xmath13 is not positively linearly independent . on a final note for this section ,
it is clear that the maximum cardinalities derived above can be achieved , as constructing an example with @xmath74 is trivial .
our first example provides a family of positively linear independent sets whose cardinality grows exponentially with dimension .
suppose @xmath80 and there exist nonnegative numbers @xmath81 such that @xmath82 as @xmath83 , there must exist some @xmath41 with @xmath84 .
since @xmath85 , and @xmath86 , there must exist some index @xmath87 with @xmath88 and @xmath89 .
this yields the contridiction @xmath90 hence , @xmath13 is positively linearly independent .
the proof of the cardinality is trivial .
[ df : convex ] a set @xmath91 is convex , if given any @xmath92 and any @xmath93 $ ] , the point @xmath94 . +
a set @xmath91 is strictly convex , if given any @xmath92 with @xmath95 and any @xmath96 , the point @xmath97 .
suppose there exists @xmath112 . by reordering , without loss of generality ,
assume @xmath113 , so @xmath114 noting that @xmath115 for all @xmath29 , we must have @xmath116 this implies @xmath117 i.e. , @xmath118 . as @xmath119 for any @xmath120 ,
fact [ fact : interior ] implies @xmath121 , contradicting @xmath122 for all @xmath29 .
it is interesting to note that the example in the proof of theorem [ thm : rn ] does not actually require @xmath109 to be finite .
indeed , in @xmath107 it is possible to create an uncountable set of vectors @xmath13 such that @xmath123 for all @xmath124 .
moreover , while the example uses the unit sphere to create the set , it is clear that @xmath108 could be replaced by the boundary of any compact strictly convex set .
this research was partially funded by the natural sciences and engineering research council ( nserc ) of canada , discover grant # 355571 - 2013 , and by the pacific institute for the mathematical sciences ( pims ) , `` optimization : theory , algorithms , and applications '' collaborative research group .
rockafellar and r.j .- b wets . .
grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , 317 .
springer - verlag berlin heidelberg 1998 , corrected 3rd printing , 2009 . | positive bases , which play a key role in understanding derivative free optimization methods that use a direct search framework , are positive spanning sets that are positively linearly independent . the cardinality of a positive basis in @xmath0 has been established to be between @xmath1 and @xmath2 ( with both extremes existing ) .
the lower bound is immediate from being a positive spanning set , while the upper bound uses _ both _ positive spanning and positively linearly independent . in this note , we provide details proving that a positively linearly independent set in @xmath0 for @xmath3 has at most @xmath2 elements , but a positively linearly independent set in @xmath0 for @xmath4 can have an arbitrary number of elements .
* keywords : * positive linear independence , positive basis , convex analysis |
about 84@xmath1 of the non - relativistic matter in the universe is in the form of cold dark matter ( cdm)@xcite , whose nature is still to be discovered .
one of the most promising hypothetical candidates that could account for the cdm observed is the axion @xcite .
the properties of the axion as the cdm particle have been studied in various papers @xcite .
we examine the possibility that the axion accounts for 100@xmath1 cdm in the light of the wmap5 mission , baryon acoustic oscillations ( bao ) and supernovae ( sn ) data .
we also upgrade the treatment of anharmonicities in the axion potential , which we find important in a specific region of the parameter space .
the axion parameter space is described by three quantities , the pq energy scale @xmath5 , the hubble parameter at the end of inflation @xmath6 and the axion initial misalignment angle @xmath3 .
the solution to the strong - cp problem by peccei and quinn @xcite introduces a new @xmath7 symmetry in the theory of strong interactions that spontaneously breaks at a temperature of order the pq energy scale , @xmath8 .
the axion is the pseudo nambu - goldstone boson originating from the breaking of @xmath7 .
we denote by @xmath9 the axion field and @xmath10 the misalignment field . at a temperature of order the qcd transition @xmath11
an effective potential @xmath12 arises through instanton effects @xcite given by @xmath13 the temperature - dependent axion mass @xmath14 is given by @xmath15 we set @xmath16mev , @xmath17 @xcite . setting @xmath18 ,
the axion mass at zero temperature is @xmath19 the integer @xmath20 represents the @xmath7 color anomaly index ; here @xmath21 .
setting @xmath22 , axion cosmology is different if the pq symmetry breaks after inflation , @xmath23 ( scenario i ) , or during inflation , @xmath24 ( scenario ii ) . over a hubble volume
, @xmath3 takes different values in scenario i , while in scenario ii @xmath3 takes only one random value .
the subsequent evolution of the zero mode of the dynamical field @xmath25 in a flat friedmann - robertson - walker metric is given by @xmath26 here a dot indicates a derivative with respect to time and @xmath27 is the hubble expansion rate .
for @xmath28 the misalignment field is frozen at a constant value , @xmath29 , where @xmath3 is the initial misalignment angle . for @xmath30
the axion field begins to oscillate in the @xmath12 potential .
the axion energy density - to - critical density is computed as @xmath31 with @xmath32 .
the function @xmath33 accounts for anharmonicities in the axion potential , i.e. , deviations from a pure harmonic potential when @xmath3 is no longer smaller than one .
we look for an analytic function @xmath33 that interpolates between previous numerical results @xcite .
we take @xmath34^{7/6}.\ ] ] the function @xmath33 in eq.([f_theta ] ) shows the correct asymptotic behaviors for @xmath35 and @xmath36 .
figure 1 shows the region of the axion parameter space in which the axion is 100@xmath1 cdm , or @xmath37 , where @xmath38 is the cdm density in units of the critical density .
the pq scale @xmath5 is bounded from below by measurements on white dwarfs cooling time , @xmath39gev , while the hubble rate at the end of inflation @xmath6 is bounded from above by non - detection of tensor modes , @xmath40gev . for @xmath23 ( scenario i ) ,
the axion can be 100@xmath1 cdm only if @xmath41 .
the function @xmath33 in eq.([f_theta ] ) affects the results by increasing the value of @xmath42 . in scenario
ii , @xmath24 , the parameter space on the top left is bounded by non - detection on axion isocurvature fluctuations that yield @xmath43 .
the upper kink in the leftmost boundary is due to the change in the dependence of the axion mass with temperature , eq.([axion_mass ] ) , while the lower kink is due to the presence of anharmonicities , eq .
( [ f_theta ] ) . in the allowed region of scenario ii , for a given @xmath5 the axion is 100@xmath1 cdm only if a correspondent value @xmath44 is selected .
this value of @xmath3 does not depend on @xmath6 .
we show the function @xmath45 in figure 2 , where we see that a value @xmath46 for a given @xmath5 makes the axion population sub - dominant compared to the cdm observed , while a value @xmath47 is over abundant .
we included the region excluded by admx1 axion direct search and future reaches by the planck satellite and future cavity searches .
50 e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys .
j. suppl .
* 180 * , 330 ( 2009 ) .
s. weinberg , phys .
lett . * 40 * , 223 ( 1978 ) ; f. wilczek , phys .
lett . * 40 * , 279 ( 1978 ) .
r. d. peccei , h. r. quinn , phys .
lett . * 38 * , 1440 ( 1977 ) .
j. preskill , m. wise , f. wilczek , phys .
b * 120 * , 127 ( 1983 ) ; l. abbott , p. sikivie , _ ibid .
_ , 133 ; m. dine , w. fishler , _ ibid .
_ , 137 ( 1983 )
. m. beltran _ et al .
_ , phys .
d * 75 * 103507 ( 2007 ) ; m. p. hertzberg _ et al .
_ , phys . rev .
d * 78 * , 083507 ( 2008 ) .
d. j. gross , r. d. pisarski , l. g. yaffe , rev .
mod . phys .
* 53 * , 43 ( 1981 ) .
m. s. turner , phys .
d * 33 * , 889 ( 1986 ) ; d. h. lyth , phys .
d * 45 * , 3394 ( 1992 ) .
necessary for the axion to be 100% of the cold dark matter in scenario ii ( @xmath48 ) , as a function of @xmath5 .
above the curve , @xmath49 .
for @xmath50gev , @xmath51 .
+ + + + + + + + + + + + , width=321 ] necessary for the axion to be 100% of the cold dark matter in scenario ii ( @xmath48 ) , as a function of @xmath5 .
above the curve , @xmath49 . for @xmath50gev , @xmath51 .
+ + + + + + + + + + + + , width=264,height=245 ] | we study for what specific values of the theoretical parameters the axion can form the totality of cold dark matter .
we examine the allowed axion parameter region in the light of recent data collected by the wmap5 mission plus baryon acoustic oscillations and supernovae @xcite , and assume an inflationary scenario and standard cosmology .
we also upgrade the treatment of anharmonicities in the axion potential , which we find important in certain cases .
if the peccei - quinn symmetry is restored after inflation , we recover the usual relation between axion mass and density , so that an axion mass @xmath0 makes the axion 100@xmath1 of the cold dark matter .
if the peccei - quinn symmetry is broken during inflation , the axion can instead be 100@xmath1 of the cold dark matter for @xmath2 provided a specific value of the initial misalignment angle @xmath3 is chosen in correspondence to a given value of its mass @xmath4 .
large values of the peccei - quinn symmetry breaking scale correspond to small , perhaps uncomfortably small , values of the initial misalignment angle @xmath3 . |
the continuum emission of many active galactic nuclei ( agn ) is variable by a factor of a few .
the broad - line region ( blr ) emission - line response to changes in the ionizing continuum provides us with important information on the size and physical properties of the emission - line clouds ( see peterson 2007 for a review ) .
a few exceptional cases of non - repetitive , i.e. transient , extreme broad - line variability have been reported , including the temporal appearance of an unusually broad heii line in ngc5548 ( peterson & ferland 1986 ) , and of double - peaked balmer lines in ngc1097 ( storchi - bergmann et al .
1993 ) and a few other liners ; interpreted as a response to transient accretion events .
detection of strong continuum and emission - line variability is of great interest , since it places tight constraints on accretion physics and physical conditions in the gas clouds surrounding the supermassive black hole ( smbh ) .
high - amplitude variability of emission lines other than the classical broad lines is rare in agn .
the one notable exception is the seyfert galaxy ic3599 .
x - ray bright during the rosat all - sky survey , a near - simultaneous optical spectrum showed strong high - ionization lines ( brandt et al .
1995 ) which were significantly fainter in a new optical spectrum taken 8 months later ( grupe et al .
the x - ray flux and the high - ionization emission lines continued to fade in subsequent years , and in low - state , the galaxy can be classified as a seyfert 1.9 or 2 .
the x - ray flare of this active galaxy was interpreted in terms of a temporarily enhanced accretion rate ( brandt et al .
1995 , grupe et al .
1995 , komossa & bade 1999 ) .
the highest amplitudes of x - ray variability , up to factors of @xmath06000 , have been almost exclusively detected in non - active galaxies , and these represent the best observational evidence to date for the process of tidal disruption of stars by smbhs ( e.g. , komossa & bade 1999 , halpern et al .
2004 , komossa et al .
2004 ) . here
we report the discovery of very unusual and variable emission lines of the galaxy sdssj0952 + 2143 at redshift @xmath1=0.079 , found in a systematic search for emission - line agn in sdss - dr6 ( sloan digital sky survey data release 6 ; adelman - mccarthy et al . 2008 ) . the optical spectrum of sdssj0952 + 2143 ( fig .
1 ) looked very similar , but more extreme , than that of ic3599 in high - state , which is why we suspected a similar mechanism at work and initiated the follow - up observations reported in this letter .
sdssj0952 + 2143 turns out to be unique in its broad multi - wavelength continuum and emission - line response to a high - energy outburst .
we use a cosmology with @xmath2=70 kms@xmath3mpc@xmath3 , @xmath4=0.3 and @xmath5=0.7 throughout this letter .
sdssj0952 + 2143 was brightest during sdss photometry performed on 2004 december 20 . at that time it showed a flat spectral energy distribution ( sed ) with an increase toward the nir ; inconsistent with the extrapolation of 2mass data taken in january 1998 ( fig .
2 ) . the sdss photometric data can be well described by a constant underlying galaxy sed and a powerlaw component consistent in shape with that during the sdss spectroscopy 1 year later , but of greater strength .
the increase towards the nir can be described by an extra black body component with a temperature of 2400 k. spectroscopy of sdssj0952 + 2143 was performed on 2005 december 30 .
it was fainter than during the photometry .
the continuum sed can be decomposed into a host - galaxy contribution , and a powerlaw of index @xmath6 , where @xmath7 . the bulk of the variability between dec .
2004 and dec .
2005 can be attributed to a variable powerlaw component which was fainter by a factor of 2 in 2005 . during the 2005 spectroscopy ,
sdssj0952 + 2143 displayed a rich and unusual emission - line spectrum .
a variety of high - ionization iron coronal lines from [ fevii ] and up to [ fexiv ] can be identified , several with unprecedented strengths relative to [ oiii]5007 , and several as strong as [ oiii ] ( fig . 1 and tab .
transitions include [ fevii]3586,3759 , rarely identified in agn .
iron lines consist of a broad ( fwhm @xmath8 kms@xmath3 ) and a narrow ( fwhm @xmath9 kms@xmath3 ) component .
strong heii4686 emission is present with a width similar to the broad component of the iron lines ( fwhm(heii)=810 kms@xmath3 ) .
the balmer line profile shows multiple components .
a multi - gaussian decomposition reveals a broad base ( fwhm(h@xmath10)=1930 kms@xmath3 , where the index b denotes the broad component of the profile ) , a narrow core ( fwhm fixed to that of [ sii]6716,31 ; 200 kms@xmath3 ) , and two remarkable unresolved narrow horns ( fig .
1 ) . the broad base is redshifted by 570 kms@xmath3 and has a luminosity of @xmath11 ergs@xmath3 .
the two narrow horns are also prominent in h@xmath12 , while its broad component is barely detected , implying a ratio h@xmath10/h@xmath13=9.5 . using stellar absorption lines detected in the spectrum
, we have determined the stellar velocity dispersion , @xmath14 kms@xmath3 .
this implies a black hole mass of @xmath15 m@xmath16 ( ferrarese & ford 2005 )
. the galex ( martin et al . 2005 ) uv photometric data points , of 2006 march 2 , are above the expected host - galaxy contribution , and are consistent with the extrapolated powerlaw component measured during sdss spectroscopy @xmath02 months earlier ( fig .
all these results locate the maximum of the outburst after the 2mass observation in 1998 , and possibly close to the 2004 sdss photometry high - state . in order to search for line variability ,
we have taken new optical spectra on 2007 december 4 and 5 with the omr spectrograph equipped with a 200 / mm grating at the xinglong 2.16 m telescope . in these new optical spectra
most of the iron lines , the heii line , and the broad h@xmath17 line are significantly fainter _ relative to [ oiii ] _
broad h@xmath17 , heii , and [ fexiv ] have faded by a factor of @xmath02 .
[ fex ] is no longer detected . despite the lower s / n of the xinglong spectrum and the coincidence of telluric atmospheric absorption with the redshifted [ fex ]
, this still implies line variability by a least a factor of two .
new photometry was performed with grond , a multi - channel imager ( greiner et al . 2008 ) attached to the 2.2 m telescope at la silla in the sdss filters g , r , i , and z on 2008 january 1 , at a seeing of 1@xmath18 .
these data confirm the decline in the optical and nir band .
the core of sdssj0952 + 2143 ( the central 1.4@xmath19 ) was a factor 2 weaker than during the 2004 sdss photometry .
we have initiated a _
chandra _ ddt x - ray observation of 10 ks duration which was carried out on 2008 february 5 .
sdssj0952 + 2143 was detected with an acis - s countrate of 0.0007 cts / s , corresponding to a ( 210)kev x - ray luminosity of 410@xmath20 erg / s ( s. komossa et al .
, in preparation ) .
we have detected the emission - line light echo and the low - energy ( nuv , optical , nir ) sed tail of a high - energy ( euv , x - ray ) outburst of the galaxy sdssj0952 + 2143 .
the euv x - ray part of the flare was not observed directly , but the presence of a strong ionizing continuum at high - state is implied by the strong h@xmath17 and heii lines and by the large iron - line fluxes relative to [ oiii ] .
further , the ionization potentials of heii and of the iron lines ( up to 358 ev ; fexiv ) imply that the ionizing continuum extended into the soft x - ray regime .
while x - ray variability is common in agn , amplitudes exceeding a factor of @xmath03050 are rare ( e.g. , komossa 2002 ) and they almost never come with a reported measurable emission - line response
. outbursts of high amplitude are of special interest , because they provide tight constraints on accretion physics under extreme conditions .
furthermore , as the flare travels through the circumnuclear material , its emission - line light echo puts us in the rare situation to be able to perform `` reverberation mapping '' not only of the blr , but potentially also of the accretion disk region , the coronal - line region ( clr ) , the molecular torus , and the narrow - line region ( nlr ) of active galaxies ( ulmer 1999 , komossa & bade 1999 ) , and it allows us to detect which of these regions are also present in non - active galaxies ( komossa 2002 ) . since x - ray monitoring simultaneous with and after
the sdss observations does not exist , we can only indirectly estimate the amplitude of variability of the outburst .
the fluxes of the iron lines , relative to [ oiii ] , observed in 2005 ( 1 year after the highest observed state in 2004 ) , are much higher than those in any other agn in sdss - dr6 . the observed ( 2005 ; post - high - state ) ratio [ fex]/[oiii ] is a factor of 60 ( 20 ) higher than the seyfert 2 ( seyfert 1 ) average of nagao et al .
( 2000 ; their tab .
6 ) , implying a high amplitude of variability .
it is possible that all emission lines were excited by the flare . alternatively ,
sdssj0952 + 2143 may show permanent low - level activity traced by low - ionization lines .
the ratio [ oiii]/h@xmath21=2.5 puts the galaxy at the border between agn and liner .
while the variable optical continuum likely is the tail of the high - energy continuum , the variable nir continuum during the 2004 sdss photometry high - state is best fit by an independent black body component of @xmath22 k , perhaps a signal of dust reprocessing before destruction .
the measured temperature slightly exceeds the sublimation temperature @xmath23 of standard silicates and graphite ( e.g. , granato & danese 1994 ) , but non - standard grain sizes and decomposition may raise the limit on @xmath23 . for the first time , transitions of [ fevii]3586,3759 as strong as [ oiii ] have been identified in a galaxy spectrum .
several [ fevii ] line ratios can be used to estimate gas temperature and gas density
. the observed high - state ratio of [ fevii]5158/[fevii]6087 @xmath24 0.2 implies a density of @xmath25 .
the ratio [ fevii]3759/[fevii]6087=1 implies a temperature of at least @xmath26 k , and up to @xmath27 k if @xmath28 ( keenan & norrington 1987 ) . the high density and temperature inferred and the variability of the iron lines on the timescale of years implies an emission - region close to the active nucleus .
possible sites of origin of the observed lines are the inner wall of the dusty torus , and/or the tidal debris of a disrupted star flung out from the system , and interacting with its environment ( khoklov & melia 1996 ) .
monitoring of the fading lines will provide a key diagnostic of the physical conditions in the line - emitting region ; if this region corresponds to the dusty torus , we can do `` torus reverberation mapping '' this way .
the h@xmath17 and h@xmath12 lines show a complex profile which can be decomposed into at least four components : a broad component with a width of 1930 kms@xmath3 redshifted by an extra 570 kms@xmath3 , a narrow component at the same redshift as other narrow forbidden lines , and two narrow horns , the more conspicuous one redshifted by 540 kms@xmath3 .
double - peaked balmer lines have been detected in a number of agn ( see eracleous et al .
2006 for a review ) . however , generally , it is the _ broad _ component of the balmer lines which is double peaked .
the strong narrow horn has no counterpart in any other emission line ; it is only detected in h@xmath17 and h@xmath12 .
one possible explanation is that it arises in a region of very high density ( such as the stellar post - disruption debris , or an accretion disk ; see sect .
3.5 ) where forbidden lines are suppressed .
the summed redshifted broad and narrow component of the h@xmath17 profile resembles to some extent , but not in detail , relativistic fek@xmath17 line profiles seen in some agn x - ray spectra ( review by miller 2007 ) .
the broad balmer decrement , h@xmath17/h@xmath12=9.5 , likely implies a significant contribution from collisional excitation indicative of high density , or else is affected by optical depth effects .
the highest amplitudes of variability detected to date , up to a factor of @xmath06000 , have essentially all been observed from non - active galaxies , and have been interpreted in terms of tidal disruptions of stars by supermassive black holes ( komossa & bade 1999 , halpern et al .
2004 , komossa et al .
2004 , and references therein ) .
stars approaching a smbh will be tidally disrupted once the tidal forces of the smbh exceed the star s self - gravity , and part of the stellar debris will be accreted , producing a luminous flare of radiation which lasts on the timescale of months to years ( e.g. , carter & luminet 1982 , rees 1988 ; see komossa 2002 for a much longer list of references on theoretical aspects of tidal disruption ) .
the outburst mechanism of sdssj0952 + 2143 is either related to processes in the accretion disk such as an instability ( only possible if sdssj0952 + 2143 is permanently active ) , or is else related to the tidal disruption of a star by the smbh at its center ( possible in both cases , active or non - active galaxy ) ; see komossa & bade ( 1999 ) for a discussion of relevant timescales of both mechanisms .
at present , we can not tell whether sdssj0952 + 2143 harbors a permanent low - luminosity agn , traced by [ oii ] and [ oiii ] , or whether these lines were also excited by the flare and represent low - density ism / nlr emission close to our line of sight . in the context of tidal disruption
, the unusual profile and double - peakedness of the balmer lines can be understood as emission from the accretion disk formed from the stellar debris after disruption .
while balmer line emission from different streams of the disrupted star looks complex during the first several months after disruption ( bogdanovic et al . 2004 ) , it would appear more like a normal disk - line thereafter . while automized procedures are currently being set up to identify new flares rapidly in the course of current x - ray surveys ( yuan et al .
2006 , esquej et al .
2007 ) , spectroscopic surveys such as sdss - ii will allow us to perform independent flare searches by the detection of variable emission - line signatures .
sdssj0952 + 2143 is exceptional among the few known flaring galaxies in its strong multi - wavelength variability , encompassing the nir , optical and nuv continuum sed , allowed and forbidden optical and nir emission lines , and the euv and x - ray sed .
such a broadband line and continuum response has not been detected before .
multi - wavelength monitoring of objects such as sdssj0952 + 2143 will enable us to follow the fading of the emission lines , providing a unique light echo mapping of the circumnuclear material potentially including the accretion disk , blr , torus , and ism ; and flung - out post - disruption stellar material if the cause of the flare was stellar tidal disruption . | we report the discovery of superstrong , fading , high - ionization iron line emission in the galaxy sdssj095209.56 + 214313.3 ( sdssj0952 + 2143 hereafter ) , which must have been caused by an x - ray outburst of large amplitude .
sdssj0952 + 2143 is unique in its strong multiwavelength variability ; such a broadband emission - line and continuum response has not been observed before .
the strong iron line emission is accompanied by unusual balmer line emission with a broad base , narrow core and double - peaked narrow horns , and strong heii emission .
these lines , while strong in the sdss spectrum taken in 2005 , have faded away significantly in new spectra taken in december 2007 .
comparison of sdss , 2mass , galex and follow - up grond photometry reveals variability in the nuv , optical and nir band . taken together , these unusual observations can be explained by a giant outburst in the euv x - ray band , detected even in the optical and nir .
the intense and variable iron , helium and balmer lines represent the `` light echo '' of the flare , as it traveled through circumnuclear material .
the outburst may have been caused by the tidal disruption of a star by a supermassive black hole .
spectroscopic surveys such as sdss are well suited to detect emission - line light echoes of such rare flare events .
reverberation - mapping of these light echoes can then be used as a new and efficient probe of the physical conditions in the circumnuclear material in non - active or active galaxies . |
the development of high - power lasers has made it possible to study the relativistic regime of laser - plasma interactions .
this is of fundamental importance for laser - induced heating and inertial confinement fusion @xcite and poses also the possibility to test fundamental physics .
the first theoretical studies of relativistic waves in plasmas date back to the fifties @xcite , where it was recognized that the the quivering velocity of electrons may lead to relativistic mass increase in an ultra - intense electromagnetic wave .
the relativistic effects can result in self - modulation and self - focusing of electromagnetic waves in plasmas @xcite .
the instability of relativistically large amplitude electromagnetic waves has been studied for magnetized electron plasmas @xcite , electron - ion plasmas @xcite , electron beam plasma systems @xcite , hot electron - ion plasmas @xcite , and magnetized hot plasmas @xcite .
the instability of relativistically strong laser light in an unmagnetized plasma has then been revisited @xcite .
the presence of multiple laser beams in a plasma can give rise to a new set of interesting phenomena @xcite .
an important application of two interacting laser beams in plasmas is the excitation of large amplitude langmuir waves , which in turn accelerate electrons to ultra - relativistic speeds @xcite .
the nonlinear coupling between two electromagnetic waves in plasmas can be described by a system of coupled nonlinear schrdinger equations that model nonlinear interactions between localized light @xcite and langmuir or ion - acoustic waves . at strongly relativistic intensities ,
laser beams can give rise to fast plasma waves via higher - order nonlinearities @xcite , or via the beat wave excitation at frequencies different from the electron plasma frequency @xcite .
particle - in - cell simulations @xcite show that large - amplitude electron plasma waves can be excited by colliding laser pulses , or by two co - propagating electromagnetic pulses where a long trailing pulse is modulated efficiently by the periodic plasma wake behind the heading short laser pulse @xcite .
the effects on parametric instabilities of a partially incoherent pump wave ( with a distribution of wave modes ) was investigated both theoretically @xcite and experimentally @xcite , where it was found that the effect of finite bandwidth is , in general , to increase the instability thresholds and lower the growth rate .
the nonlinear interaction between two electromagnetic waves in the earth s ionosphere has also been considered @xcite . in that case , it is , in addition , important to focus attention on the collisional coupling between the waves , e.g. @xcite .
related phenomena occur also in semiconductor plasmas @xcite .
our previous treatment of the instability of two coupled laser beams in an electron - ion plasma @xcite and a two - temperature electron plasma @xcite were limited to the investigation of the relativistic raman and brillouin scattering instabilities in one dimension , and two - dimensional effects were only partly included .
the purpose of the present paper is to include the multi - dimensional effects , which are particularly important for the cases where an electromagnetic wave decays into one low - frequency wave and one high - frequency electrostatic wave that are propagating obliquely to the pump wave .
let us consider the propagation of intense laser light in an electron ion plasma .
the dynamics of the high frequency laser light is governed by @xmath0 where @xmath1 is the electron plasma frequency , @xmath2 is the equilibrium number density , @xmath3 is the magnitude of the electron charge , @xmath4 is the electron mass , @xmath5 is the electron thermal speed , @xmath6 is the speed of light , and the normalized slow time - scale electron number density perturbation is @xmath7 .
the latter is excited by the ponderomotive force of the high frequency waves .
if the ions are considered as immobile , we have @xcite @xmath8 whereas if the electrons are treated as inertialess @xcite @xmath9 where @xmath10 is the sound speed .
we will here consider two large amplitude electromagnetic waves @xmath11 .
we then have @xmath12 .
using equations ( 1)(3 ) we can now derive a nonlinear dispersion relation .
the calculations are straightforward but lengthy .
following closely the analysis of ref .
@xcite , it then turns out to be convenient to introduce a characteristic velocity @xmath13 that is defined by @xmath14\ ] ] where @xmath15 , @xmath16 , @xmath17 and @xmath18 is the pump frequency and wavevector of the pump wave @xmath19 ( @xmath20 ) , whereas @xmath21 and @xmath22 are associated with the electrostatic low - frequency fluctuations .
furthermore , we have here introduced the pump wave quiver velocity @xmath23 , @xmath24 where @xmath25 represents the sideband damping @xcite , and @xmath26 . for the pump frequencies , we use @xmath27 . in the present paper
we will , for simplicity , further assume that @xmath28 .
this means that double resonance parametric phenomena @xcite , where the difference between the frequencies @xmath29 and @xmath30 is close to twice a natural frequency , will be neglected . choosing two transverse waves that propagate in the @xmath31 and @xmath32-directions , respectively with the pump velocities in the @xmath33 and @xmath34-directions , respectively
, we note that ( 4 ) reduces to @xmath35\ ] ] with these limitations , and following refs . @xcite and @xcite , the nonlinear dispersion relation turns out to be @xmath36 & & \quad + \frac{|\mathbf{k}\cdot\mathbf{a}_{01}|^2}{{k}_{1+}^2\omega_{1+}^2\varepsilon_{1 + } } + \frac{|\mathbf{k}\cdot\mathbf{a}_{01}|^2}{{k}_{1-}^2\omega_{1-}^2\varepsilon_{1- } } + \frac{|\mathbf{k}\cdot\mathbf{a}_{02}|^2}{{k}_{2+}^2\omega_{2+}^2\varepsilon_{2 + } } + \frac{|\mathbf{k}\cdot\mathbf{a}_{02}|^2}{{k}_{2-}^2\omega_{2-}^2\varepsilon_{2- } } = 0 .
\label{dispersion}\end{aligned}\ ] ] in eq .
( 6 ) , as well as below , we have normalized @xmath37 by @xmath38 . for electron langmuir waves
, we have @xcite @xmath39 whereas for ion acoustic waves @xcite @xmath40
) of the two - plasmon decay of one laser beam for different electron thermal speeds a ) @xmath41 , b ) @xmath42 , c ) @xmath43 , and d ) @xmath44 .
the pump amplitude is @xmath45 , the pump wavevector @xmath46 .
the second laser beam intensity is set to zero ( @xmath47 ) .
, width=377 ] ) of the two - plasmon decay of one laser beam for wavevectors a ) @xmath48 , b ) @xmath46 , c ) @xmath49 , and d ) @xmath50 .
the pump amplitude is @xmath45 , and the electron thermal speed @xmath44 .
the second laser beam intensity is set to zero ( @xmath47 ) .
, width=377 ] ) of the two - plasmon decay of two coupled laser beams for wavevectors a ) @xmath48 and @xmath51 , b ) @xmath46 and @xmath52 , c ) @xmath49 and @xmath53 , and d ) @xmath49 and @xmath52 .
the pump amplitudes are @xmath45 and @xmath54 , and the electron thermal speed is @xmath44 .
, width=377 ] ) of the ion parametric decay instability of one laser beam for different electron thermal speeds a ) @xmath41 , b ) @xmath42 , c ) @xmath43 , and d ) @xmath44 ; in each case the ion acoustic speed is set to @xmath55 .
the pump amplitude is @xmath45 , and the pump wavevector is @xmath56 .
the second laser beam intensity is set to zero ( @xmath47 ) .
, width=377 ] ) of the ion parametric decay instability of one laser beam for wavevectors a ) @xmath57 , b ) @xmath56 , c ) @xmath58 , and d ) @xmath59 .
the pump amplitude is @xmath45 , the electron thermal speed is @xmath44 , and the ion acoustic speed is @xmath60 .
the second laser beam intensity is set to zero ( @xmath47 ) .
, width=377 ] ) of the ion parametric decay instability of two coupled laser beams for wavevectors a ) @xmath56 and @xmath61 , b ) @xmath58 and @xmath62 , c ) @xmath59 and @xmath63 , and d ) @xmath59 and @xmath62 .
the pump amplitudes are @xmath45 and @xmath54 , the electron thermal speed is @xmath44 and the ion acoustic speed is @xmath60 .
, width=377 ] we have carried out a numerical analysis of eq .
( [ dispersion ] ) , where we have assumed that the frequency is complex valued , where the imaginary part of @xmath21 represents the growth rate . in our treatment , we have concentrated on three - wave decay processes where the important terms in eq .
( [ dispersion ] ) are the ones with down - shifted daughter waves .
hence , we have kept the resonant terms with subscripts `` - '' in eq .
( [ dispersion ] ) but neglected those with subscripts `` + '' . for large wavenumbers , we recover almost identically the stimulated raman and brillouin cases treated in ref .
thus , we shall here concentrate on the regimes of smaller wavenumbers in which the two - plasmon decay and ion parametric decay instabilities become important , and where it is crucial to have a fully multi - dimensional treatment .
we first consider the two - plasmon case in which the low - frequency wave is a langmuir wave , and where @xmath64 in eq .
( [ dispersion ] ) equals @xmath65 in ( [ langmuir ] ) .
the growth rates ( normalized by @xmath66 ) are presented in figs .
we have here denoted the unit vectors in the @xmath34 and @xmath32 directions by @xmath67 and @xmath68 . in fig .
1 , the electron thermal effects are investigated .
the dispersion of the langmuir waves due to the electron pressure causes the instability to restricted in a bounded domain in wavevector space , with a maximum growth rate in a direction perpendicular to the propagation direction of the laser beam . in fig .
2 , we have investigated how the instability depends on the pump wavenumber of a single electromagnetic beam .
we find that the instability is dominant for wavenumbers equal to or larger than @xmath69 , and that the instability vanishes for smaller wavenumbers of the pump wave .
the maximum instability occurs for wavevectors almost perpendicular to the laser beam propagation direction .
the instability maximum occurs for wavenumbers much larger than the pump wavenumber .
eventually kinetic effects will become important and electron landau damping will decrease the growth rate .
the case of two coupled electromagnetic beams is investigated in fig .
3 . while beam 1 is directed in the @xmath34-direction as in figs . 1 and 2
, beam @xmath70 is here in the @xmath32-direction , perpendicular to beam 1 .
we see in panel a ) of fig .
3 that the coupled beam system gives rise to a new instability with a maximum growth rate in the direction of the dichotome in the center between the two beam propagation directions . for wavenumbers larger than @xmath69 , the instability becomes more evenly distributed in all directions , but with a well - defined maximum growth rate for some wavenumber . for the case where the wavenumber of beam 2 is smaller than the one of beam 1 , in panel d ) , we see a superposition of the two instability regions for the separate beams . we note that similar effects can also appear due to direct subharmonic wave generation in nonuniform plasmas @xcite .
we next investigate the parametric decay instability in which the electromagnetic wave decays into one electrostatic wave and one low - frequency ion acoustic wave , where @xmath64 in eq .
( [ dispersion ] ) equals @xmath71 in ( [ ion ] ) .
the growth rates are presented in figs .
( 4)(6 ) .
we have here concentrated on the long wavelength limit where the main instability is the ion parametric decay instability , in which the electromagnetic wave decays into one slightly frequency - downshifted electrostatic wave and one low - frequency ion acoustic wave . in fig .
4 , we have studied the thermal effect on the instability of one single electromagnetic beam .
we see that for higher electron thermal and ion acoustic speeds , the region of instability becomes smaller in wavenumber space , and that there is a well - defined maximum of the instability for propagation almost perpendicular to the wave propagation direction . in fig . 5 , we have considered different wavenumbers of the pump wave .
we see that for large wavenumbers , the maximum instability occurs for larger wavevectors perpendicular to the pump wavevector .
finally , we consider the interaction between two coupled electromagnetic beams in fig .
6 . for the cases of equal pump amplitudes and lengths of the wavevectors , in panels a)c )
, the instability becomes almost rotationally symmetric , with equal maximum growth rates in all directions .
for the case where the wavenumber of beam 2 is smaller than the one of beam 1 , in panel d ) , we see a superposition of the two instability regions for the separate beams . for the ion parametric decay instability
, we could not see the same distinct amplification of the instabilities as we could see for the two - plasmon decay in panel a ) of fig .
3 . in conclusion
, we have investigated the instability of two coupled large - amplitude electromagnetic waves in a plasma .
our investigation shows that two - plasmon decay plays an important role for pump frequencies approximately two times larger than the background electron plasma frequency , and that one has a maximum growth rate perpendicular to the propagation direction of the electromagnetic wave . for two coupled electromagnetic beams
, there is here a new and stronger instability in a direction between the two laser beams . the ion parametric decay instability , in which an electromagnetic wave decays into one langmuir wave and one ion acoustic wave , is important for long wavelengths of the electromagnetic pump wave , where the instability leads to ion acoustic waves that are propagating perpendicularly to the electromagnetic beam direction . here
, the addition of a second electromagnetic beam leads to a superposition of the instabilities .
however , we see no strong nonlinear amplification due to the two beams . our study could be important for both laser - plasma interactions and for ionospheric heating experiments @xcite .
kodama , r. _ et al .
_ 2001 _ nature _ * 412 * , 798 + kodama , r. _ et al .
_ 2002 _ nature _ * 418 * , 933 .
akhiezer , a. i. , and polovin , r. v. 1956 _ zh . eksp .
fiz . _ * 30 * , 915 .
[ 1956 _ sov .
phys . jetp _ * 3 * , 696 . ] max , c. e. , arons , j. , and langdon , a. b. 1974 _ phys .
_ * 33 * , 209 .
stenflo , l. 1976 _ phys .
_ * 14 * , 320 .
+ stenflo , l. 1981 _ phys .
* 23 * , 779 .
stenflo , l. and shukla , p. k. 1984 _ phys . rev .
* 30 * , 2110 .
shukla , p. k. , yu , m. y. , and stenflo , l. 1986 _ phys .
a _ * 34 * , 1582 .
tsintsadze , n. l. and stenflo , l. 1974 _ phys .
a _ * 48 * , 399 .
tsintsadze , n. l. , tskhakaya , d. d. , and stenflo , l. 1979 _ phys .
a _ * 72 * , 115 .
tsintsadze , n. l. , papuashvili , n. a. , tsikarishvili , e. c. , and stenflo , l. 1980 _ phys .
_ * 21 * , 183 .
shukla , p. k. , rao , n. n. , yu , m. y. , and tsintsadze , n. l. 1986 _ phys .
_ * 138 * , 1 .
sakharov , a. s. and kirsanov , v. i. 1994 _ phys .
e _ * 49 * , 3274 .
+ quesnel , b. , mora , p. , adam , j. c. , gurin , s. _ et al .
_ 1997 _ phys .
lett . _ * 78 * , 2132 .
+ barr , h. c. , mason , p. , and parr , d. m. 1999 _ phys .
_ * 83 * , 1606 .
rosenbluth , m. n. and liu , c. s. 1972 _ phys .
* 29 * , 701 . | the three - dimensional instability of two coupled electromagnetic waves in an unmagnetized plasma is investigated theoretically and numerically . in the regime of two - plasmon decay , where one pump wave frequency is approximately twice the electron plasma frequency ,
we find that the coupled pump waves give rise to enhanced instability with wave vectors between those of the two beams . in the case of ion parametric decay instability , where the pump wave decays into one langmuir wave and one ion acoustic wave ,
the instability regions are added with no distinct amplification .
our investigation can be useful in interpreting laser - plasma as well as ionospheric heating experiments . |
in @xcite , zelevinsky has introduced an involution of the grothendieck group of the category of complex smooth representations of finite length of @xmath4 , where @xmath5 is a @xmath6-adic field , and conjectured that this involution permutes the classes of irreducible representations ( @xcite , 9.17 ) . in @xcite , zelevinsky further conjectured a geometric description of this involution in terms of the graded nilpotent orbits that parametrize the simple @xmath7-modules .
both conjectures have been proved by moeglin and waldspurger @xcite for the category of admissible representations of @xmath7 generated by their space of @xmath8-fixed vectors , where @xmath8 is an iwahori subgroup of @xmath7 . in this case , by a theorem of bernstein , borel and matsumoto , the conjectures can be reformulated in terms of the hecke algebra @xmath9 of @xmath7 with respect to @xmath8 .
this is an affine hecke algebra of type @xmath10 with parameter @xmath11 equal to the cardinality of the residue field of @xmath5 .
as shown in @xcite , the problem then becomes that of describing how the isomorphism classes of simple @xmath9-modules are permuted when the action is twisted by a certain involutive automorphism @xmath12 of @xmath9 .
in fact the answer does not depend on @xmath11 , and is the same for any complex parameter @xmath1 of infinite multiplicative order .
this is no longer the case if one considers the hecke algebra @xmath0 with @xmath1 a primitive @xmath2th root of 1 .
according to a conjecture of vigneras @xcite , the solution of the problem in this case would be relevant for the @xmath13-modular representation theory of @xmath7 , where @xmath13 is a prime different from @xmath14 , and @xmath2 is the multiplicative order of @xmath11 in the field with @xmath13 elements . in this paper , we obtain results similar to those of moeglin and waldspurger in the root of unity case .
namely , ( i ) given a simple @xmath0-module @xmath15 , we describe in section 5.2 a simple combinatorial algorithm for computing the isomorphism class @xmath16 $ ] of the simple module obtained by twisting with @xmath12 , and ( ii ) we prove in section 5.3 that the geometric description conjectured by zelevinsky still holds in this case , this time in terms of the @xmath3-graded nilpotent orbits parametrizing the simple @xmath0-modules . however , our methods are different from those of @xcite ( and thus we obtain new proofs of these results in the non root of unity case ) .
our main tool is a recent theorem of ariki @xcite which relates the grothendieck groups of the algebras @xmath0 with the enveloping algebra @xmath17 , where @xmath18 is the kac - moody algebra of type @xmath19 if @xmath1 has order @xmath2 , and @xmath20 if @xmath1 has infinite order . using this result
, we show that our problem is equivalent to describing the natural 2-fold symmetry of kashiwara s crystal graph of @xmath21 , induced by the root diagram automorphism exchanging the simple roots @xmath22 and @xmath23 .
then the combinatorial algorithm for @xmath16 $ ] follows immediately from the explicit description of this graph ( theorem [ th1 ] ) , and the equivalent formulation in terms of nilpotent orbits is deduced from a geometric construction of the crystal basis conjectured by lusztig @xcite and recently verified by kashiwara and saito @xcite .
we note that procter @xcite and aubert @xcite have established independently that the first conjecture of zelevinsky holds for the whole category of complex smooth representations of finite length of @xmath7 . assuming this hypothesis , moeglin and walspurger had already shown in @xcite that their geometric and combinatorial descriptions of zelevinsky s duality were also valid for this larger class of representations .
the paper is structured as follows . in section [ sect2 ]
we recall the definition of zelevinsky s involution , and the parametrization of simple @xmath0-modules by multisegments
. then we can formulate our problem in a more precise way . in section [ sect3 ]
we review the hall - ringel algebra @xmath24 associated with a quiver and the canonical basis of its subalgebra @xmath25 . in section [ sect4 ] , using the isomorphism @xmath26 , we compute the action of the chevalley generators of @xmath21 on the pbw - basis of @xmath24 .
this allows us to describe the crystal graph of @xmath21 in lusztig s parametrization by nilpotent orbits , and to derive the combinatorial description of @xmath12 ( 5.2 ) .
finally , in 5.3 we briefly recall the results of lusztig , kashiwara and saito on the construction of the crystal basis in terms of irreducible components of a certain lagrangian variety , and we deduce from them the geometric description of @xmath12 .
[ [ section ] ] let @xmath1 be a nonzero complex number .
the affine hecke algebra @xmath0 associated to @xmath27 is the associative @xmath28-algebra with invertible generators @xmath29 and @xmath30 subject to the relations @xmath31 the algebra @xmath0 admits involutive automorphisms @xmath12 , @xmath32 , @xmath33 , given on the generators by @xmath34 one checks that @xmath12 and @xmath32 commute and that @xmath35 .
the involution @xmath33 was defined by iwahori and matsumoto in @xcite , p. 280 , and @xmath12 is usually called zelevinsky s involution ( see @xcite , p. 146 ) , because of its relation with the involution on representations of @xmath6-adic @xmath27 introduced by zelevinsky in @xcite .
[ [ section-1 ] ] for @xmath36 a composition of @xmath37 , set @xmath38 and denote by @xmath39 the subalgebra of @xmath0 generated by @xmath30 and @xmath40 . for @xmath41 ,
let @xmath42 be the 1-dimensional representation of @xmath39 defined by @xmath43 , @xmath44 , and @xmath45 , @xmath46 , where @xmath47 .
we denote by @xmath48 the induced @xmath0-module @xmath49 let @xmath50 be the complexified grothendieck group of the category of finite dimensional @xmath0-modules , and let @xmath51 be the linear span of the classes of the composition factors of the various @xmath48 in @xmath50 .
we set @xmath52 where we have put for convenience @xmath53 . when @xmath54 is a permutation of @xmath55 , @xmath56 and @xmath57 for some @xmath58 , the induced modules @xmath48 and @xmath59 are in general non isomorphic , but their classes in @xmath60 are equal @xcite ( see @xcite , p. 455 ) . in the sequel
, we shall always work in @xmath60 and therefore shall only consider pairs @xmath55 up to permutation .
in fact , such unordered pairs @xmath55 are naturally identified with certain graded nilpotent orbits @xmath61 ( see below section [ sect3 ] ) , which gives a canonical labelling of the classes of the induced modules by these orbits : @xmath62 = [ m_o]$ ] .
there is a partial order on orbits given by @xmath63 if @xmath64 .
when @xmath1 is not a root of unity , it is known @xcite that for all @xmath61 , there is a unique simple module whose class occurs in the expansion of @xmath65 $ ] but does not occur in any @xmath66 $ ] with @xmath67 .
let @xmath68 denote this simple module .
then the @xmath68 are pairwise non isomorphic , and one has in @xmath60 @xmath69 = [ l_o ] + \sum_{o\lhd p } k_{o , p } \ , [ l_p]\ ] ] for some positive integers @xmath70 .
( the @xmath70 are in fact values at 1 of kazhdan - lusztig polynomials of type @xmath10 , as conjectured by zelevinsky @xcite and proved by ginzburg @xcite , theorem 8.6.23 . )
if @xmath1 is a primitive @xmath2th root of unity , we may consider that in a pair @xmath55 , @xmath71 belongs to @xmath72 .
let us say that @xmath55 is aperiodic if for each @xmath73 , the set @xmath74 has at most @xmath75 elements .
these are the labels associated to the so - called aperiodic graded nilpotent orbits ( see @xcite , 15.3 ) .
it follows from ariki s theorem @xcite that for all aperiodic @xmath61 , there is a unique simple module @xmath68 whose class occurs in the expansion of @xmath65 $ ] but does not occur in any @xmath66 $ ] with @xmath67 .
moreover , @xmath76 \
o \mbox { aperiodic } \}$ ] is a basis of @xmath60 .
finally , the multiplicities of the @xmath77 $ ] in the @xmath65 $ ] are given by certain kazhdan - lusztig polynomials of affine type @xmath78 .
( for the general type a similar result was previously announced by grojnowski @xcite ) .
[ [ section-2 ] ] when @xmath1 is not a root of unity , the pairs @xmath55 can be identified with zelevinsky s multisegments @xcite .
we recall that a segment is an interval @xmath79 $ ] in @xmath80 , and that a multisegment is a formal finite unordered sum @xmath81 $ ] .
( here @xmath82 stands for the multiplicity of the segment @xmath79 $ ] in @xmath83 . )
the multisegment corresponding to @xmath55 is then @xmath84\,.\ ] ] a multisegment can be conveniently regarded as a coloured multipartition , as a sequence @xmath85 of partitions , such that the parts of @xmath86 are the lengths of the segments @xmath79 $ ] , and each cell of the @xmath87th column of the young diagram of @xmath86 contains the integer @xmath88 .
in the case where @xmath1 is a primitive @xmath2th root of unity , we shall also identify the labels @xmath55 with multisegments , this time over @xmath3 , where we regard @xmath3 as a set of @xmath2 points on the unit circle .
for @xmath89 and @xmath90 we define the segment of length @xmath91 and origin @xmath92 on this circle ( or more appropriately the loop ) as : @xmath93\,.\ ] ] sometimes we shall also need the dual notation @xmath94 : = [ i-\ell+1 , i-\ell+2 , \ldots , i]\,.\ ] ] a multisegment over @xmath3 is now a formal finite sum @xmath95 } \ , ( k;j]\,.\ ] ] the cyclic multisegment corresponding to the label @xmath55 is @xmath96
. these cyclic multisegments may also be represented by @xmath2-tuples of partitions @xmath97 , with cells labelled by integers modulo @xmath2 .
a label @xmath55 is aperiodic if for each @xmath91 , there is at least one component @xmath86 which does not contain the part @xmath91 ( this is , up to conjugation of partitions , the labelling used by ringel in @xcite for a basis of the composition algebra of the cyclic quiver ) .
these definitions are illustrated in figure [ fig1 ] .
let @xmath98 ( @xmath3 ) .
the set of all multisegments over @xmath8 will be denoted by @xmath99 , or simply @xmath100 if no confusion can arise .
thus we have a parametrization of the simple @xmath0-modules @xmath101 by @xmath102 ( with @xmath83 aperiodic for @xmath103 ) .
we introduce a @xmath104-grading of @xmath60 by putting @xmath105 = ( d_i)_{i\in i}$ ] , where @xmath106 is the number of cells of the diagram of @xmath83 which contain @xmath92 .
[ [ section-3 ] ] the three involutions @xmath12 , @xmath32 and @xmath33 induce involutions on the set of simple @xmath0-modules @xmath101 , and hence on the set of multisegments @xmath83 that parametrize them .
it is easy to see that @xmath107 where if @xmath108 , then @xmath109 $ ] .
therefore , it is equivalent to describe either @xmath12 or @xmath33 . for generic @xmath1 ,
a geometric description of @xmath12 in terms of graded nilpotent orbits has been conjectured by zelevinsky @xcite .
this conjecture was proved by moeglin and waldspurger @xcite , who also gave a combinatorial algorithm for computing @xmath110 . an explicit formula for @xmath111
has also been found by knight and zelevinsky @xcite .
the finite hecke algebra @xmath112 is the quotient of @xmath0 by the relation @xmath113 .
hence , @xmath33 induces an involution on simple @xmath112-modules .
when @xmath1 is not a root of unity , these modules are parametrized by partitions of @xmath37 , and @xmath33 is just the conjugation of partitions .
when @xmath1 is an @xmath2th root of unity , the simple @xmath112-modules are labelled by @xmath2-regular partitions , partitions with no part repeated more than @xmath75 times , and it was recently proved that @xmath33 coincides with a bijection previously defined by mullineux @xcite .
one proof follows from @xcite p. 229 and
ariki s theorem , and a second proof was obtained by brundan @xcite by quantizing " the results of kleshchev on modular representations of symmetric groups .
the aim of this paper is to describe the maps @xmath12 and @xmath33 in the affine case when @xmath1 is a primitive @xmath2th root of unity .
proposition [ pro5.1 ] below can be regarded as a common generalization of mullineux s bijection as described by kleshchev @xcite , and of the algorithm of moeglin and waldspurger .
[ [ section-4 ] ] let @xmath114 denote the quiver of type @xmath115 ( @xmath20 or @xmath19 ) .
the set of vertices of @xmath114 is @xmath116 $ ] ( @xmath80 or @xmath3 ) , and the set of arrows is @xmath117 .
a nilpotent representation of @xmath114 over the field @xmath118 is a pair @xmath119 , where @xmath120 is an @xmath8-graded finite - dimensional vector space , and @xmath121 is a degree -1 nilpotent linear operator on @xmath122 . in other words
, @xmath123 is an element of @xmath124 , the subset of nilpotent elements of @xmath125 the vector @xmath126 is the ( graded ) dimension of the representation .
the group @xmath127 acts on @xmath128 and @xmath124 by conjugation .
two representations @xmath119 and @xmath129 are equivalent if @xmath123 and @xmath130 lie in the same orbit .
thus , the set of isomorphism classes of nilpotent @xmath114-representations is naturally identified with @xmath131 , where @xmath132 is the set of @xmath133-orbits in @xmath124 for any graded @xmath118-vector space @xmath122 of dimension @xmath134 .
the isomorphism classes of nilpotent @xmath114-representations are parametrized by @xmath99 in the following way . for each vertex @xmath135 , let @xmath136 $ ] denote the simple @xmath114-module for which @xmath137 and @xmath138 .
given a positive integer @xmath91 , there is a unique ( up to isomorphism ) indecomposable @xmath114-module @xmath139 $ ] of length @xmath91 with head @xmath136 $ ] , and all indecomposable modules are of this type .
therefore , any nilpotent @xmath114-module @xmath140 is isomorphic to @xmath141 : = \bigoplus_{i\in i , \ell \in \nb^ * } \k[\ell;i]^{\oplus m_{(\ell;i]}}\ ] ] for a unique multisegment @xmath83 .
we denote by @xmath142 the @xmath133-orbit of the module @xmath143 $ ] . for simplicity , we also denote by @xmath144 the orbit of the simple module @xmath136 $ ] .
[ [ section-5 ] ] we recall the definition of the twisted hall algebra associated with the quiver @xmath114 ( see @xcite ) .
the classification of nilpotent @xmath114-modules is independent of the ground field .
take @xmath145 , the field with @xmath11 elements , and let @xmath146 be orbits in @xmath147 respectively , where @xmath148 .
let @xmath140 be a @xmath114-module lying on the orbit @xmath149 .
then it is was proved by ringel @xcite that the number of submodules of @xmath140 with type @xmath150 and cotype @xmath61 is a polynomial in @xmath11 with integer coefficients , independent of the choice of @xmath140 on @xmath149 .
this is the hall polynomial @xmath151 .
define a bilinear form @xmath37 on @xmath104 by @xmath152 then the generic twisted hall algebra associated with @xmath114 is the @xmath153-algebra @xmath24 with basis @xmath154 and multiplication @xmath155 ( in the setting of @xcite , @xmath156 stands for the product of characteristic functions of the sets of @xmath157-rational points of @xmath61 for a fixed prime @xmath14 . ) [ [ section-6 ] ] let @xmath61 be a @xmath133-orbit in @xmath124 .
one sets @xmath158 where @xmath159 denotes the dimension of the orbit @xmath61 ( not to be confused with the dimension @xmath160 of the representation of @xmath114 corresponding to a point of @xmath61 ) . the basis @xmath161 is called the poincar - birkoff - witt basis of @xmath24 .
if @xmath150 is a @xmath162-orbit in @xmath163 , it follows from ( [ mult1 ] ) that @xmath164 where @xmath165 in the sequel , we shall use another expression for @xmath166 . for @xmath167 ,
set @xmath168 and denote by @xmath169 the dimension of the space of endomorphisms of a module @xmath140 lying in @xmath61 .
then @xmath170 indeed , @xmath171 and @xmath172 thus , if we set @xmath173 and @xmath174 , we have @xmath175 @xmath176 [ [ section-7 ] ] let @xmath177 be the twisted composition algebra of @xmath114 , _
i.e. _ the subalgebra of @xmath178 generated by the characteristic functions @xmath179 of the orbits of the simple modules @xmath136 $ ] .
when @xmath114 is of type @xmath115 or @xmath20 , one has @xmath180 , but @xmath25 is strictly contained in @xmath24 for the type @xmath19 .
the vector space @xmath24 has a basis @xmath181 given by @xmath182 where @xmath183 is the stalk at a point of @xmath184 of the @xmath92th intersection cohomology sheaf of the closure @xmath185 of @xmath61 . for @xmath114 of type @xmath20
, these sheaves have been first considered by zelevinsky in @xcite , where he made a conjecture about their connection with representations of @xmath6-adic @xmath186 known as the @xmath6-adic analogue of the kazhdan - lusztig conjecture . in @xcite
, zelevinsky further proved that the poincar polynomials @xmath187 are equal to certain kazhdan - lusztig polynomials @xmath188 of type @xmath10 . for @xmath114 of type @xmath19
, lusztig has shown that the poincar polynomials are also kazhdan - lusztig polynomials , this time of affine type @xmath78 ( @xcite,11 ) . in @xcite , he introduced the subfamily of aperiodic nilpotent orbits .
these are the orbits @xmath142 parametrized by aperiodic multisegments ( see 2.3 ) .
it is proved in @xcite , section 5 that @xmath189 is a basis of @xmath25 .
note that non aperiodic orbits @xmath184 may occur in the right - hand side of ( [ cano ] ) for an aperiodic orbit @xmath61 .
let @xmath190 be the ring automorphism of @xmath25 defined by @xmath191 and @xmath192 .
let @xmath193 be the @xmath194$]-submodule of @xmath24 spanned by the pbw basis @xmath195 .
let @xmath196 and suppose that @xmath61 is aperiodic if @xmath114 is of type @xmath19 .
then @xmath197 is characterized as the unique element of @xmath198 such that @xmath199 and @xmath200 .
[ [ section-8 ] ] let @xmath18 be the kac - moody lie algebra associated with the quiver @xmath114 of type @xmath115 ( @xmath20 or @xmath19 ) , namely , @xmath201 ( @xmath202 or @xmath203 ) .
let @xmath204 be the corresponding quantized universal enveloping algebra , with generators @xmath205 .
the subalgebra generated by @xmath206 is denoted by @xmath21 .
this algebra is isomorphic to the twisted composition algebra @xmath25 of @xmath114 , the isomorphism being given by @xmath207 .
it follows that the twisted hall algebra @xmath24 can be regarded as a left @xmath21-module .
the basis @xmath208 of @xmath25 becomes via this isomorphism a basis of @xmath21 : this is lusztig s canonical basis .
it is known that it coincides with kashiwara s lower global basis of @xmath21 . by taking @xmath209 one
obtains the canonical basis of @xmath17 , that we shall still denote by @xmath208 .
the aim of this section is to describe kashiwara s crystal graph of @xmath24 .
to do so , we shall first obtain explicit formulas for the action of the chevalley generators @xmath210 and their adjoint @xmath211 on @xmath24 .
we believe that these formulas are of independent interest .
for example they yield an algorithm similar to that of @xcite for computing the canonical basis in its expansion on the pbw - basis .
[ [ section-9 ] ] we describe the action of the @xmath210 on @xmath24 in the pbw - basis .
we assume from now on that @xmath114 is of type @xmath19 .
( the other two cases are readily deduced from this one , since @xmath212 is a natural subalgebra of @xmath213 , and @xmath214 is the limit @xmath215 of @xmath212 " ) . as explained above , the pbw - basis is naturally labelled by multisegments @xmath83 over @xmath3 . for simplicity ,
we abuse notation and write @xmath216 for the corresponding element of the pbw - basis . given @xmath217 , @xmath89 , and a multisegment @xmath83 such that @xmath218}\not = 0 $ ]
if @xmath219 , we define a new multisegment @xmath220 & \hbox { if $ \ell=1 $ , } \cr & \m + ( \ell;i ] - ( \ell-1;i-1 ] & \hbox { if $ \ell>1$. } } \right.\ ] ] in case @xmath218 } = 0 $ ] , we put @xmath221 .
[ pro4.1 ] the chevalley generators of @xmath222 act on the pbw - basis of @xmath24 by @xmath223 } - m_{(k;i ] } ) } \,[m_{(\ell;i ] } + 1 ] \ ,
\<\m^+_{\,\ell , i}\ > \,,\ ] ] where for @xmath224 , @xmath225 = ( v^a - v^{-a})/(v - v^{-1})$ ] is the usual @xmath226-integer .
we have to apply formulas ( [ mult2 ] ) and ( [ twistr ] ) with @xmath227 and @xmath228 .
we see easily that here , @xmath229 , and @xmath230 } \,.\ ] ] also , it is clear that in this case , the orbits @xmath149 giving a nonzero contribution to the right - hand side of ( [ mult2 ] ) are those parametrized by the multisegments of the form @xmath231 for some @xmath73 .
we distinguish two cases depending on whether @xmath232 or @xmath233 . in the first case we have @xmath234 ) ) + \dim(\hom ( \kb[i],\kb[\m ] ) ) + \dim ( \hom ( \kb[\m],\kb[i ] ) ) \\[2 mm ] & = & 1 + \sum_{k\in \nb^ * } m_{[i;k ) } + \sum_{k\in \nb^ * } m_{(k;i]}\,.\end{aligned}\ ] ] to calculate the hall polynomial @xmath235 , let us count the number of submodules @xmath119 of @xmath236 $ ] isomorphic to @xmath237 $ ] . for @xmath238 we must have @xmath239 , and @xmath240 must be a hyperplane in @xmath241 which ( i ) contains @xmath242 ( so that @xmath122 be stable under @xmath130 ) and ( ii ) does not contain @xmath243 ( so that @xmath119 be isomorphic to @xmath237 $ ] ) .
the number of hyperplanes satisfying ( i ) is clearly @xmath244 } { ] \hspace*{-.5mm}]}$ ] and the number of those satisfying ( i ) but not ( ii ) is @xmath245 } { ] \hspace*{-.5mm}]}$ ] .
( here we have used the notation @xmath246\hspace*{-.5mm}]}= ( q^a-1)/(q-1)$ ] . ) therefore , @xmath247 } { ] \hspace*{-.5mm}]}- { [ \hspace*{-.5mm}[}\sum_{k\ge 2 } m_{(k;i ] } { ] \hspace*{-.5mm}]}= q^{\sum_{k\ge 2 } m_{(k;i]}}\ , { [ \hspace*{-.5mm}[}1 + m_{(1;i ] } { ] \hspace*{-.5mm}]}\,,\ ] ] and finally , the coefficient of @xmath248 in @xmath249 is equal to @xmath250 } -\sum_{k\in \nb^ * } m_{[i;k ) } + \sum_{k\in \nb^ * } m_{(k;i-1 ] } -2\sum_{k\ge 2 } m_{(k;i ] } -m_{(1;i ] } } [ m_{(1;i ] } + 1 ] \ ] ] @xmath251 } - m_{(k;i ] } ) } \,[m_{(1,i ] } + 1]\,.\ ] ] in the case @xmath219 , the calculation is similar and one gets @xmath252 } - \sum_{k\le \ell-1 } m_{(k;i-1]}\ , , \\[2 mm ] f_{o , p}^q(q ) & = & { [ \hspace*{-.5mm}[}1 + \sum_{k\ge \ell } m_{(k;i ] } { ] \hspace*{-.5mm}]}- { [ \hspace*{-.5mm}[}\sum_{k\ge \ell + 1 } m_{(k;i ] } { ] \hspace*{-.5mm}]}= q^{\sum_{k\ge \ell + 1 } m_{(k;i]}}\ , { [ \hspace*{-.5mm}[}1 + m_{(\ell;i ] } { ] \hspace*{-.5mm}]}\,,\end{aligned}\ ] ] so that the coefficient of @xmath253 in @xmath249 is equal to @xmath254 } - \sum_{k\le \ell-1 } m_{(k;i-1 ] } -\sum_{k\in \nb^ * } m_{[i;k ) } + \sum_{k\in \nb^ * } m_{(k;i-1 ] } } \qquad \qquad\ ] ] @xmath255}}\ , v^{-m_{(\ell;i]}}\ , [ m_{(\ell;i ] } + 1 ] \ ] ] @xmath256 } - m_{(k;i ] } ) } \,[m_{(\ell , i ] } + 1]\,.\ ] ] * remark * the specialization @xmath209 of proposition [ pro4.1 ] appears in @xcite lemma 4.2 .
[ [ section-10 ] ] to define the kashiwara operators @xmath257 and @xmath258 of the @xmath222-module @xmath24 , one needs to introduce endomorphisms @xmath211 of the space @xmath24 ( see @xcite , 3.4 ) .
first one defines a scalar product on @xmath24 for which the basis @xmath259 is orthogonal by @xmath260 } \over \left |\aut \fq [ \m ] \right |_{q = v^{-2 } } } \ , \delta_{\m\,,\ , \m ' } \,.\ ] ] here we have written for short @xmath261 .
this scalar product is the one introduced by green @xcite , multiplied by a normalization factor @xmath262}$ ] so that it induces kashiwara s scalar product @xcite on @xmath263 .
[ lem4.1 ] put @xmath264 .
we have @xmath265 \right | = \left |\end \fq [ \m ] \right | \ , \prod_{i,\ell } { \left |\gl(m_{(\ell ; i]},\fq ) \right | \over \left |\mat(m_{(\ell ; i]},\fq ) \right | } = q^{\varepsilon(\m)}\ , \prod_{i,\ell } \varphi_{m_{(\ell;i]}}(q^{-1 } ) \,.\ ] ] let @xmath119 be a @xmath114-module isomorphic to @xmath237 $ ] , and let @xmath266 denote the subspace of @xmath240 consisting of all the generators of an indecomposable summand of @xmath119 of type @xmath267 $ ] .
thus , @xmath268}$ ] .
an endomorphism @xmath269 of @xmath119 is completely determined by its restriction @xmath270 to @xmath271 .
let @xmath272 be a complement of @xmath273 in @xmath240 , and @xmath274 be the corresponding projection of @xmath240 onto @xmath266 .
then , @xmath275 implies that @xmath276 for @xmath277 . on the other hand , @xmath278 is not submitted to any condition and may be an arbitrary element of @xmath279},\fq)$ ] .
now , @xmath269 is an automorphism of @xmath119 if and only if @xmath280 is in @xmath281 , where @xmath282 .
hence @xmath269 is an automorphism if and only if @xmath278 belongs to @xmath283},\fq)$ ] , and the lemma follows . using lemma [ lem4.1 ] and the formula
= \sum_{i,\ell } \ell \ , m_{(\ell ; i ] } $ ] , we rewrite ( [ eq22 ] ) as @xmath285 } \choose 2 } } \ , { ( 1-v^2)^{(\ell-1)m_{(\ell;i ] } } \over [ m_{(\ell;i ] } ] ! } \ , \delta_{\m\,,\ , \m ' } \,.\ ] ] next we define @xmath286 as the adjoint operator of @xmath210 with respect to this scalar product .
the following proposition describes the action of @xmath211 on the pbw - basis .
given @xmath217 , @xmath89 , and a multisegment @xmath83 such that @xmath287}\not = 0 $ ] , we put @xmath288 & \hbox { if $ \ell=1 $ , } \cr & \m - ( \ell;i ] + ( \ell-1;i-1 ] & \hbox { if $ \ell>1$. } } \right.\ ] ] in case @xmath287 } = 0 $ ] , we put @xmath289 .
[ pro4.2 ] the endomorphisms @xmath211 act on the pbw - basis of @xmath24 by @xmath290 } - m_{(k;i ] } ) } \,v^{-m_{(1;i ] } + 1 } \,\<\m^-_{\,1,i}\ > \qquad \qquad \qquad
\qquad \qquad \qquad\ ] ] @xmath291 } - m_{(k;i ] } ) } \,v^{-m_{(\ell;i ] } + 1 } \,(1-v^{2(m_{(\ell-1;i-1]}+1 ) } ) \,\<\m^-_{\,\ell , i}\ > \,.\ ] ] by definition of @xmath211 and using proposition [ pro4.1 ] , the coefficient of @xmath292 in @xmath293 is equal to @xmath294 } - m_{(k;i ] } ) } \,[m_{(\ell;i]}]\,.\ ] ] now by ( [ sp ] ) , if @xmath219 @xmath295 } + m_{(\ell-1;i-1 ] } + 1 } \ , { [ m_{(\ell-1;i-1 ] } + 1 ] \over [ m_{(\ell;i ] } ] } \,,\ ] ] and if @xmath296 @xmath297 } + 1 } \ , \over [ m_{(1;i ] } ] } \,,\ ] ] which gives the required result .
* remark * propositions [ pro4.1 ] , [ pro4.2 ] should be compared to the formulas of @xcite for the action of the chevalley generators of @xmath213 on the level @xmath13 fock spaces .
actually , our formulas can be regarded as the limit @xmath298 of those of @xcite .
[ [ section-11 ] ] let @xmath299 be kashiwara s crystal graph of @xmath24 for type @xmath19 in lusztig s geometric parametrization .
that is , the vertices of this graph are the multisegments @xmath83 ( or the corresponding orbits @xmath142 ) , and there is an arrow @xmath300 in the graph if and only if @xmath301 , where @xmath257 denotes as usual the kashiwara operator @xcite . in particular , the connected component @xmath302 of this graph containing the empty multisegment gives the crystal graph of @xmath263 parametrized by the set of aperiodic orbits .
there are several descriptions of the crystal graph of @xmath21 available in the literature ( see @xcite , @xcite , @xcite ) , but the only ones using lusztig s parametrization are given by reineke @xcite and kashiwara - saito @xcite .
unfortunately , reineke s description is restricted to finite dimensional @xmath18 , while kashiwara - saito s description is in terms of the geometry of orbits , and does not seem to yield a combinatorial algorithm ( see section 5.3 ) .
our description of @xmath299 will be an extension to type @xmath19 of reineke s result for type @xmath115 .
[ th1 ] let @xmath83 be a multisegment over @xmath3 , and fix @xmath217 .
for @xmath303 , put @xmath304 } - m_{(\ell;i]})$ ] and let @xmath305 be minimal such that @xmath306 .
then there is an arrow @xmath307 in the crystal graph @xmath299 of @xmath24 .
our strategy is to reduce the problem to reineke s description of the crystal graph of @xmath308 .
we first note that one can let @xmath2 tend to infinity in reineke s work and get the crystal graph of @xmath309 .
let us identify @xmath217 with its representative in @xmath310 .
choose @xmath2 integers @xmath311 such that @xmath312 and @xmath313 .
for example @xmath314 if @xmath315 , and @xmath316 if @xmath317 .
define an embedding @xmath318 of the set of @xmath3-multisegments in the set of @xmath80-multisegments by @xmath319}\,(\ell;r]\right ) = \sum_{\ell \in \nb^*}\sum_{r\in \zb / n\zb } m_{(\ell;r]}\,(\ell;j_r]\,.\ ] ] then define a @xmath153-linear embedding @xmath320 by @xmath321 let @xmath322 denote the chevalley element in @xmath309 and @xmath323 its adjoint .
proposition [ pro4.1 ] and [ pro4.2 ] readily imply [ lemg ] ( i ) the subspace @xmath324 of @xmath309 is stable under @xmath322 and @xmath323 .
@xmath325 let @xmath326 ( @xmath327 ) be the subgraph of the crystal of @xmath24 ( @xmath309 ) obtained by erasing all arrows labelled by @xmath328 .
lemma [ lemg ] shows that @xmath329 thus , @xmath318 induces an isomorphism of graphs between @xmath326 and the full subgraph of @xmath327 with vertex set @xmath330 , where @xmath83 runs over all @xmath3-multisegments .
the description of this last graph follows from @xcite and this finishes the proof of theorem [ th1 ] .
theorem [ th1 ] is illustrated in figure [ fig2 ] in the case @xmath331 .
* remarks 1 * it follows easily from theorem [ th1 ] that the crystal @xmath299 of @xmath24 decomposes into an infinite number of connected components isomorphic to @xmath302 , the highest weight vertices of these components being the periodic multisegments , that is , the multisegments @xmath83 for which @xmath332}=m_{(\ell , s]}$ ] for all @xmath333 .
this is in agreement with @xcite , remarks 3.4.10 , 3.5.1 . * 2 * the crystal graphs of the level @xmath13 integrable representations of @xmath203 as described by @xcite embed in a natural way into the crystal graph of @xmath24 ( 4.3 remark ) .
each @xmath13-tuple of partitions @xmath334 is mapped to the multisegment obtained by taking the formal sum of the rows of the @xmath335 .
[ [ section-12 ] ] let @xmath336 denote the graded dual of the @xmath104-graded space @xmath60 .
this is made into an associative algebra by taking as product the dual of the comultiplication @xmath337 coming from the restriction maps @xmath338 for any @xmath339 let @xmath340 be the one - dimensional module such that @xmath341 .
if @xmath1 is generic ( @xmath342 ) then @xmath343 ( @xmath344 ) is a basis of @xmath345 .
the dual basis vectors are denoted by @xmath346 .
recall that if @xmath140 is a simple @xmath347-module then @xmath348 acts as a scalar , say @xmath349 , on @xmath140 .
then the @xmath92-restriction of the class @xmath350 $ ] is defined in @xcite as the class @xmath92-@xmath351 $ ] of the @xmath352-submodule @xmath353 and @xmath92-@xmath354 is extended to @xmath60 by linearity .
it is straightforward to check that for @xmath355 and @xmath356 , one has @xmath357 [ lem5.1 ] ( i ) @xmath33 induces an automorphism of @xmath336 such that @xmath358.(ii ) @xmath12 induces an anti - automorphism of @xmath336 such that @xmath359 . claim ( i ) is obvious since @xmath33 commutes with the natural embedding @xmath360 claim ( ii ) follows from the identity @xmath361 where @xmath362 , @xmath363 , and @xmath364 stands for the natural embedding of @xmath365 into @xmath347 .
[ [ section-13 ] ] it was proved by ariki that the map @xmath366 is an isomorphism of algebras @xmath367 if @xmath342 ( @xmath368 if @xmath1 is generic ) ( @xcite , prop .
4.3 ; see also the appendix ) .
recall from 2.2 that @xmath76\}$ ] is a basis of @xmath60 .
ariki s theorem states that under the previous isomorphism , one has @xmath369^*$ ] , that is , the canonical basis of @xmath370 ( @xmath371 ) is mapped to the basis of @xmath336 dual to @xmath76\}$ ] .
[ pro5.1 ] let @xmath83 be a @xmath80-multisegment ( an aperiodic /n - multisegment ) and let @xmath372 be generic ( @xmath342 ) .
let @xmath373 be the corresponding simple @xmath0-module .
then , the twisted module @xmath374 is isomorphic to @xmath375 , where @xmath376 is the vertex of the crystal graph of @xmath377 ( @xmath222 ) obtained from the vertex @xmath83 by the 2-fold symmetry @xmath378 of the graph .
let @xmath33 denote the automorphism of @xmath17 induced by the @xmath379-fold symmetry @xmath378 of the dynkin diagram of @xmath18 . repeating the argument of @xcite 7.3 , 7.4
, we see that @xmath33 preserves the canonical basis @xmath380 and maps @xmath381 to @xmath382 , where @xmath376 is obtained from @xmath83 by the symmetry @xmath378 of the crystal graph .
the result then follows from lemma [ lem5.1 ] ( i ) and ariki s theorem . using theorem [ th1 ] , it is then easy to describe a simple algorithm for computing @xmath376 .
take any path @xmath383 from the empty multisegment to @xmath83 in the crystal graph .
then @xmath384 is a path to @xmath376 .
this is illustrated in figure [ fig3 ] in the case @xmath331 .
recall that zelevinsky s involution @xmath12 is related to @xmath33 by @xmath385 . as explained in 2.4 , @xmath32 acts on a multisegment @xmath83 by simply changing all segments @xmath267
$ ] into @xmath386 .
thus proposition [ pro5.1 ] gives as well an algorithm for @xmath12 .
* remarks 1 * consider the crystal of @xmath309 . among all paths joining a given @xmath83 to @xmath387
, there is a distinguished one obtained by iterating the following procedure .
take the segments @xmath388 $ ] of @xmath83 with @xmath92 minimum , and among them pick up one with @xmath91 minimum . then there is always an arrow @xmath389 in the graph .
the algorithm for @xmath12 obtained by using this particular path is precisely the one described by moeglin and waldspurger ( @xcite , lemme ii.3 ) .
note that here , we need a total order on the vertices @xmath92 of the quiver , and there is no obvious analog of this procedure for the cyclic case .
* 2 * in @xcite , berenstein and zelevinsky have described for @xmath201 the involutions of the canonical basis of @xmath17 corresponding to @xmath33 and @xmath12 via lemma [ lem5.1 ] and ariki s theorem .
they have related them to the transition map between the two indexations of the basis associated with the two extreme reduced expressions of @xmath390 ( @xcite , prop .
they also noted the coincidence with the multisegment duality coming from the representation theory of @xmath6-adic @xmath27 ( see @xcite , 3 .
remark ) .
[ [ section-14 ] ] let @xmath18 be a symmetrizable kac - moody lie algebra .
we briefly review the geometric construction of @xmath302 conjectured by lusztig ( @xcite , 10.4 ) and proved by kashiwara - saito @xcite .
let @xmath114 be the graph associated to the root system of @xmath18 .
let @xmath8 be the set of vertices .
fix an orientation @xmath391 of @xmath114 and set @xmath392 , where @xmath393 denotes the orientation opposite to @xmath391 . for
any arrow @xmath394 let @xmath395 and @xmath396 be the input and output vertices of @xmath71 .
given a @xmath8-graded complex vector space @xmath397 , lusztig has introduced @xcite @xmath398 and as above , @xmath399 the space @xmath400 is naturally identified with @xmath401 , and , thus , @xmath402 gets identified with the cotangent bundle of @xmath128 .
it is therefore a symplectic manifold , and the natural action of @xmath133 on @xmath403 is hamiltonian , with moment map @xmath404 , where @xmath405 is the lie algebra of @xmath133 .
the @xmath92-th component of @xmath406 is the map
@xmath407 where @xmath408 if @xmath409 and @xmath410 if @xmath411 . the set @xmath412
is easily seen to be the union of the conormal bundles @xmath413 of the @xmath133-orbits @xmath414 .
an element @xmath415 is nilpotent if there is an integer @xmath13 such that @xmath416 for all sequences @xmath417 such that the above expression is meaningful .
define @xmath418 it is known @xcite that @xmath419 is a closed equidimensional @xmath133-stable subvariety of @xmath403 . for @xmath114 of type @xmath115 , @xmath20 or @xmath19 , an element @xmath420 of @xmath419 can be identified with a pair @xmath421 of commuting nilpotent endomorphisms of @xmath122 of degree @xmath410 and @xmath422 respectively . moreover , the irreducible components of @xmath419 are the closures @xmath423 of the conormal bundles to the nilpotent @xmath133-orbits @xmath61 in @xmath128 for type @xmath115 or @xmath20 ( @xcite , 14 ) , and to the aperiodic nilpotent @xmath133-orbits for type @xmath19 ( @xcite , 15 ) . to the @xmath8-graded space @xmath122 we associate the negative root @xmath424
let @xmath425 be the set of irreducible components of @xmath419 .
lusztig has introduced a crystal structure on the set @xmath426 , and conjectured that this graph is isomorphic to the crystal @xmath302 of @xmath427 ( @xcite , 8.9 ) .
this was proved by kashiwara - saito @xcite .
more precisely , recall that the vertices of @xmath302 are identified with orbits @xmath196 for type @xmath115 or @xmath20 , and aperiodic orbits @xmath196 for type @xmath19 .
then the map @xmath428 is an isomorphism of crystals @xmath429 .
the algebra @xmath204 admits an anti - automorphism @xmath430 such that @xmath431 , @xmath432 , and @xmath433 , which induces an involution ( still denoted by @xmath430 ) of @xmath302 @xcite .
it is shown in @xcite , section 5.3 that the involution of @xmath425 induced by the transpose map @xmath434 on @xmath419 coincides with @xmath430 .
( this was also conjectured by lusztig @xcite , 10.4 . )
more precisely , if @xmath435 then @xmath436 is the irreducible component containing @xmath437 , where @xmath438 is generic such that @xmath439=0 $ ] . summarizing the discussion , and taking into account lemma [ lem5.1 ] ( ii ) and ariki s theorem , we can state : let @xmath114 be of type @xmath19 or @xmath20 . for @xmath123 in @xmath128 ,
set @xmath440 assume that @xmath123 is nilpotent ( and aperiodic in type @xmath19 ) .
then , ( i ) there is a unique @xmath133-orbit @xmath441 such that @xmath442 is zarisky dense and open in @xmath443 , and this orbit only depends on the orbit @xmath61 of @xmath123;(ii ) if @xmath444 and @xmath445 , then @xmath446 , where the multisegment @xmath447 is obtained from @xmath83 as explained in 5.2 . for @xmath114 of type @xmath20 , we recover the description of zelevinsky s involution conjectured in @xcite and established in @xcite .
in @xcite ariki uses the non - vanishing theorem for affine hecke algebras at roots of unity announced in @xcite . the purpose of this appendix is to indicate a simple proof of this non - vanishing result for type @xmath10 .
s. ariki has informed one of us ( e.v . ) that this simple argument was known to him and to g. lusztig . for the sake of completeness
, we include it here .
fix @xmath448 and put @xmath342 .
we use the same notations as in sections 3.1 and 5.3 .
in particular @xmath449 is the cyclic quiver with @xmath2 vertices and @xmath122 is a @xmath8-graded vector space of dimension @xmath450 .
let @xmath451 be the set of sequences @xmath452 such that @xmath453 .
given such a sequence let @xmath454 be the variety of flags of type @xmath455 in @xmath122 ( see @xcite , section 1 ) .
let @xmath456 be the variety of pairs @xmath457 such that @xmath458 is @xmath123-stable . as usual
, we denote by @xmath459 the bounded derived category of complexes of @xmath28-sheaves on a complex algebraic variety @xmath460 .
let @xmath461 be the second projection and put @xmath462 . by definition the degree-@xmath463 component of lusztig s canonical basis @xmath208 is labelled by the @xmath133-orbits @xmath464 such that a shift of the intersection cohomology complex @xmath465 is a direct factor in @xmath466 ( use @xcite , section 2.4 ) .
let @xmath5 be the variety of complete flags in @xmath467 .
the cotangent variety @xmath468 is identified with the set of pairs @xmath469 such that @xmath458 is @xmath123-stable .
if @xmath470 is semi - simple let @xmath471 be its centralizer , let @xmath472 be the set of nilpotent matrices @xmath123 such that @xmath473 and let @xmath474 be the set of pairs @xmath475 such that @xmath476 and the flag @xmath458 is fixed by @xmath477 . let @xmath478 be the second projection and set @xmath479 .
given a sequence @xmath450 such that @xmath480 , if @xmath477 is a diagonal matrix such that @xmath481 has the multiplicity @xmath106 in the spectrum of @xmath477 then @xmath482 , @xmath483 are identified with @xmath133 , @xmath124 above . in this case
set @xmath484 .
then , ginzburg s geometric construction of the simple modules of the affine hecke algebra ( see @xcite ) implies that the simple modules of @xmath485 are labelled by the @xmath133-orbits @xmath464 such that a shift of the complex @xmath465 is a direct factor in @xmath486 . in our case
the non - vanishing result is precisely that the orbits labelling the degree-@xmath37 component of the basis @xmath208 are the same as the orbits labelling the simple @xmath485-modules .
it is a consequence of the obvious equality of the complexes @xmath487 and @xmath466 ( simply observe that @xmath488 ) . * acknowledgements * we are grateful to m .- f
. vigneras for posing the question which motivated this work . b.l . and j .- y.t . would like to thank a. zelevinsky for his beautiful series of lectures on `` representations of quantum groups and piecewise linear combinatorics '' given at the universit de marne - la - valle in 1994 , which arose their interest in this field .
abc , _ on the decomposition numbers of the hecke algebra of @xmath489 _ , j. math .
kyoto univ .
* 36 * ( 1996 ) , 789808 . , _ dualit dans le groupe de grothendieck de la catgorie des reprsentations lisses de longueur finie dun groupe @xmath14-adique _ , trans .
, * 347 * ( 1995 ) , 2179 - 2189 ; erratum trans .
a.m.s . , * 348 * ( 1996 ) , 4687 - 4690 . and a. zelevinsky , _ canonical bases for the quantum group of type @xmath490 and piecewise - linear combinatorics _ , duke math
. j. * 82 * ( 1996 ) , 473 - 502 . , _ modular branching rules and the mullineux map for hecke algebras of type a _ , to appear in proc .
london math .
soc .. and v. ginzburg , _ representation theory and complex geometry _ , birkhauser 1997 . , _ hall algebras , hereditary algebras and quantum groups _ , invent . math . * 120 * ( 1995 ) , 361377 .
, _ representations of affine hecke algebras ( and affine quantum @xmath491 ) at roots of unity _ , internat .
notices * 5 * ( 1994 ) , 215217 . and
h. matsumoto , _ on some bruhat decomposition and the structure of the hecke rings of @xmath492-adic chevalley groups _
math . i.h.e.s .
* 25 * ( 1965 ) , 237280 . and
m. okado , _ combinatorics of representations of @xmath493 at @xmath494 _ , commun . math . phys .
* 136 * ( 1991 ) , 543566 . and k. misra , _ crystal bases of verma modules for quantum affine algebras _ , compositio mathematica * 92 * ( 1994 ) , 299 - 325 .
, _ on crystal bases of the @xmath11-analogue of universal enveloping algebra
_ , duke math . j.
* 63 * ( 1991 ) , 465516 . ,
_ crystal base and littlemann s refined demazure character formula _
, duke math . j.
* 71 * ( 1993 ) , 839 - 858 . and
y. saito , _ geometric construction of crystal bases _ , duke math .
j. * 89 * ( 1997 ) , 936 . , _ branching rules for the modular representations of symmetric groups iii ; some corollaries and a problem of mullineux _
, j. london math .
soc . , * 2 * ( 1995 ) . and
a. zelevinsky , _ representations of quivers of type @xmath10 and the multisegment duality _ , adv . math . * 117 * ( 1996 ) , 273293 . and j .- y .
thibon , _ hecke algebras at roots of unity and crystal bases of quantum affine algebras _ , commun .
* 181 * ( 1996 ) , 205 - 263 . , _ canonical bases arising from quantized enveloping algebras , i _ , j. amer .
soc . * 3 * ( 1990 ) , 447 - 498 . , _ canonical bases arising from quantized enveloping algebras , ii _
, progress theor .
supplement * 102 * ( 1990 ) , 175 - 201 . , _
quivers , perverse sheaves and quantized enveloping algebras _ , j. amer .
soc . * 4 * ( 1991 ) , 365 - 421 . ,
_ affine quivers and canonical bases _ , publ . math .
, i.h.e.s . * 76 * ( 1992 ) , 111 - 163 .
, _ canonical bases and hall algebras _ , lectures at the summer school on representation theory and algebraic geometry , montral 1997 . and
waldspurger , _
sur linvolution de zelevinski _
, j. reine angew . math .
* 372 * ( 1986 ) , 136177 .
, _ bijections of @xmath14-regular partitions and @xmath14-modular irreducibles of the symmetric groups _ , j. london math . soc . , * 20 * ( 1979 ) , 60 - 66 . and a. zelevinsky , _ polyhedral realizations of crystal bases for quantized kac - moody algebras _ , advances math . *
131 * ( 1997 ) , 253 - 278 . , _ the zelevinsky duality conjecture for @xmath491 _ , thesis king s college , london , 1994 .
, _ on the coloured graph structure of lusztig s canonical basis _ ,
* 307 * ( 1997 ) , 705 - 723 . , _ hall algebras _ , in _ topics in algebra _ , banach center publ .
* 26 * ( 1990 ) , 433 - 447 . , _ hall algebras revisited _ , in _ quantum deformations of algebras and their representations _ , israel math .
conf . proc . , ed .
a. joseph , s. shnider , * 7 * ( 1993 ) , 171 - 176 . , _ the composition algebra of a cyclic quiver _ , proc .
london math .
* 66 * ( 1993 ) , 507 - 537 .
, _ on modules over the hecke algebra of a @xmath14-adic group _ , invent .
* 79 * ( 1985 ) , 443 - 465 .
, _ a propos dune conjecture de langlands modulaire _ , in _ finite reductive groups : related structures and representations _ ,
cabanes , marc ( ed . ) , birkhauser 1997 . , _ induced representations of reductive @xmath14-adic groups ii _ , ann .
* 13 * ( 1980 ) , 165210 . , _ a @xmath14-adic analog of the kazhdan - lusztig conjecture _ ,
appl . * 15 * ( 1981 ) , 83 - 92 .
, _ two remarks on graded nilpotent classes _ , russian math .
surveys * 40 * ( 1985 ) , 249250 . | we give a combinatorial algorithm for computing zelevinsky s involution of the set of isomorphism classes of irreducible representations of the affine hecke algebra @xmath0 when @xmath1 is a primitive @xmath2th root of 1 .
we show that the same map can also be interpreted in terms of aperiodic nilpotent orbits of @xmath3-graded vector spaces .
epsf.tex -16 mm -20 mm = eufm10 = cmbx7 = msbm10 = 11 [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] # 1#21=1=1 1 by # 1 2=1 2 by # 1 1=1=1 by .4pt 1=1 2 by .4pt 1=2 1 # 1 2 t p 2s o # 1height 0 mm depth # 1 mm width 0 mm # 1#2height # 1 mm depth # 2 mm width 0 mm # 1#2 ( # 1 # 2 ) # 1#2_q # 1#2 ( ) 0.6 cm |
high temperature simulations are common tools for the investigation of protein unfolding properties @xcite . due to the inherent long timescales apparent in the unfolding
processes @xcite , a decrease in computation time is even nowadays still desirable . by the usage of elaborated temperatures which are typically above the protein melting temperature @xmath0 , a significant decrease of unfolding times
can be observed @xcite .
as it is often assumed , an arrhenius - type behavior of unfolding kinetics is responsible for this remarkable acceleration @xcite while the pathway of unfolding is unchanged @xcite .
+ over the years , the desirability of using elaborated temperatures has lead to several novel insights .
it has been shown that the properties of the solvent related to the chosen water model at higher temperatures are in agreement to experimental results @xcite .
hence the validity of the unfolding pathway is not influenced by spurious artefacts of the environment .
furthermore it is known that at very high simulation temperature the unfolding process occurs more rapidly due to a simultaneous breaking of several intramolecular interactions such that intermediate states can not be detected considerably @xcite . to avoid this drawback ,
the concept of an upper critical temperature has been introduced @xcite .
carrying out simulations below this temperature leads to an enhanced sampling of intermediate states such that the complete unfolding pathway can be studied in detail .
hence the transformation of high temperature to low temperature pathways is methodologically validated although a detailed comparison is often missing .
+ the application of elaborated temperature unfolding simulations can be also useful to study the unfolding mechanisms for specific dna and rna structures due to large free energy barriers .
prominent examples for these configurations are non watson - crick like structures in dna like the g - quadruplex structures and the i - motif @xcite . + the first one is formed by guanine ( g ) rich sequences @xcite while the latter is present in more cytosine ( c ) rich strands of dna @xcite . although binding seems to be unfavorable for these structures , stability is achieved by a proton mediated cytosine binding between different strands or regions of the sequence resulting in a c - ch@xmath1 pairing @xcite . hence an acidic environment is able to spend a proton which leads to a hemi - protonated cytosine mimicking an ordinary c - g binding as it is present in common dna .
it becomes clear that these structures have been found to be only stable at slightly acidic to neutral conditions resulting in ph values ranging from 4.8 to 7.0 @xcite .
furthermore i - motif structures show a remarkable stability @xcite and have been found as tetrameric and double stranded complexes between different molecules although they are also occurring in single stranded dna @xcite . a picture of the c - ch@xmath1 complex and the corresponding single stranded i - motif with its sequence
is shown in fig .
+ however , a detailed investigation of the function in the human cell is still missing although the applicability as a new class of possible targets for cancers and other diseases has been discussed @xcite . pairing ( bottom left ) which is responsible for the formation of the dna i - motif ( top left ) with the corresponding sequence ( right ) where c , t and a denote cytosine , adenine and thymine .
, scaledwidth=35.0% ] in contrast to this lack of knowledge , the usage of the i - motif in modern biotechnology has experienced an enormous growth over the last years @xcite .
since the i - motif becomes unstable at basic ph values , a systematic decrease and increase of protons in the solution by changing the ph value leads to a reversible folding and unfolding mechanism .
it has been shown that the unfolding and folding occurs on a timescale of seconds @xcite . due to these properties
the presence of i - motifs in modern nano applications is obvious .
+ technological examples for the i - motif are given by molecular nanomachines @xcite , switchable nanocontainers @xcite , sensors to detect the ph value inside living cells @xcite , building materials for logic gate devices @xcite and sensors for distinguishing single walled and multi - walled carbon nanotube systems @xcite .
+ recently it has been assumed @xcite , that the grafting density massively influences the structure of an i - motif layer due to steric hindrance in the underlying folding and unfolding pathway .
hence a detailed investigation of the folding respectively unfolding pathway especially for high grafting densities in nanodevices of the i - motif and under physiological conditions is of prior importance .
+ in this paper we present the results of high temperature 500 k molecular dynamics simulations concerning the unfolding mechanism of a maximum unstable single stranded dna i - motif structure without hemiprotonated cytosine .
our results indicate a fast initial decay of the i - motif leading to stable conformations . by detailed investigation of the essential dynamics of a biased metadynamics simulation
, we were able to identify the main unfolding pathways at 300 k. the good agreement to the high temperature unfolding pathway is validated .
our findings can be related to the calculated underlying free energy landscape which allows to identify the most stable conformations .
+ the paper is organized as follows . in the next section we present the theoretical background of the metadynamics method and the essential dynamics methodology .
then we introduce the numerical details and present the results of our simulations .
we conclude with a brief summary .
the system we consider is described by a set of coordinates @xmath2 evolving under the action of dynamics following the trajectory @xmath3 and described by a canonical equilibrium distribution at temperature @xmath4 .
the set of coordinates @xmath2 may include atomic positions or angles as well as any other auxiliary collective variable representing the characteristics of the system .
+ if the system shows metastability , some regions separated by large energy barriers can not be explored by the evolution of the trajectories for low temperatures in a reasonable simulation time . thus reflecting the original idea of metadynamics , an additional potential energy at specific constant times @xmath5
is applied on the trajectory @xmath6 to overcome the barriers and to accelerate the rare events @xcite .
typically this potential energy is given by gaussian hills which are summed at time @xmath7 to @xmath8 where @xmath9 are the @xmath10 collective variables , @xmath11 denotes the gaussian height and @xmath12 is the gaussian width . by applying this procedure subsequently ,
free energy minima can be escaped and the unfolding pathway can be identified @xcite .
furthermore it has been shown @xcite that the added potential energy resembles the underlying free energy @xmath13 after a long simulation time . to overcome certain drawbacks
@xcite , we use a recently published variant of the metadynamics method @xcite . within this method , the metadynamics potential is calculated on a grid leading to a constant calculation time in contrast to ordinary metadynamics .
the refinement of the resulting landscape is achieved by a histogram reweighting procedure of several biased simulation runs @xcite .
a projection scheme allows to change the collective variables a posteriori @xcite .
the main part of a successful calculation relies on the choice of appropriate collective variables on which the free energy landscape and the resulting unfolding pathway is spanned .
well - suited collective variables are the essential eigenvectors of the system @xcite .
it has been recently demonstrated that the application of eigenvectors as collective variables results in adequate descriptions of the free energy landscapes in ordinary metadynamics computations @xcite .
thus we follow this approach due to the assumption that nearly all relevant motion is captured in the first eigenvectors of the system @xcite . in the following
we give a brief description of the method .
+ the essential eigenvectors can be calculated by the superimposed coordinates @xmath14 of @xmath15 atoms of the system which build the covariance matrix @xmath16 via @xmath17 where @xmath18 and the brackets @xmath19 denote the reference value .
the diagonalization of @xmath16 leads to @xmath20 where @xmath21 is a matrix of eigenvectors and @xmath22 is a matrix of eigenvalues marking the positional fluctuations .
sorting the eigenvalues in decreasing order allows to identify the largest positional fluctuations by the first eigenvectors which form the essential subspace with all important structural transitions @xcite .
the projection at time @xmath7 on the @xmath23th eigenvector @xmath24 is then defined by @xmath25 with the specific eigenvectors ranging from @xmath26 @xcite .
+ to compare the motion of several subsets of eigenvectors for different trajectories , a class of functions has been introduced @xcite .
the surface overlap @xmath27 between @xmath15 eigenvectors @xmath28 from one set and @xmath29 eigenvectors from another subset @xmath30 can be calculated by the squared inner product @xmath31 which gives an estimate for the similarity of two sets .
if the spaces are completely orthogonal the relation gives @xmath32 in contrast to identity which results in @xmath33 . additionally the inner product matrix can be calculated by @xmath34 which allows to derive a detailed comparison between the motion of different subset or trajectories and the corresponding eigenvectors .
we have performed our molecular dynamics simulations of the i - motif in explicit tip3p solvent at t=300 k and t=500 k by the gromacs software package @xcite .
the single dna strand consists of 22 nucleic acid bases given by the sequence @xmath35_3-t-3^{\prime}$ ] where @xmath4 , @xmath36 and @xmath37 denote thymine , adenine and cytosine .
we modeled this structure which is directly related to the sequence used in @xcite by slight modifications of the pdb entry 1eln @xcite .
our periodic simulation box contains 5495 tip3p water molecules and 22 sodium ions to compensate the charging of the dna .
all interactions have been calculated by using the ffamber03 force field @xcite .
after energy minimization , the initial warm up phase of 1 ns has been performed by keeping the position of the i - motif restrained .
the cubic simulation box with periodic boundary conditions has a dimension of @xmath38 nm . we applied a nose - hoover thermostat to the system where all bonds have been constrained by the lincs algorithm @xcite .
electrostatics have been calculated by the pme algorithm @xcite and the time step was @xmath39 fs .
the calculation of the free energy landscape has been performed by the method presented in @xcite .
the biased metadynamics simulations at 300 k for the investigation of the unfolding pathway have been conducted by the program plug - in plumed @xcite .
the gaussian hills were set each 2 ps with a height of 0.25 kj / mol and a width of 0.25 nm .
the corresponding reaction coordinates for the biased energy are the distance between nucleobases c1 and t22 and the distance between the center - of - mass for the c1-t22 to the a11 nucleobase as they were also used in @xcite .
the final eigenvector free energy landscapes have been calculated by the projection scheme proposed in @xcite where the landscape has been refined via 15 biased simulations of 10 ns at 300 k and 500 k. a detailed protocol for the explicit calculation of the eigenvector free energy landscape in correspondence to the projection scheme is presented in @xcite . + for a detailed investigation of the unfolding mechanism we furthermore conducted five 500 k simulations and five 300 k unbiased simulations each with 10 ns duration and averaged the resulting values .
the root - mean square deviation is used for the study of the unfolding pathway @xmath40 with the positions @xmath41 of @xmath15 atoms for different times and the distances @xmath42 between two atoms where @xmath43 is the reference or average position . due to reasons of simplicity , we paired each three nucleobases in one group resulting in the sequence cys1-taa1-cys2-taa2-cys3-taa3-cys4-t .
we chose the initial structure as reference position and calculated the rmsd for the 500 k and the 300 k unbiased simulations for several groups .
we start this section by presenting the values for the root - mean square deviation of the unbiased simulations at 500 k and 300 k in fig .
[ fig2 ] .
the initial opening of the i - motif appears on a timescale up to 200 ps for the 500 k simulations .
this is obvious by regarding the results for the slightly increasing rmsd until it becomes nearly constant from 200 ps to 400 ps .
after 400 ps a stable structure can be observed with minor fluctuations indicated by a constant rmsd of the taa1/taa3 which is present up to 1.2 ns . as it has been discussed in @xcite for 300 k ,
a reason for this stable conformation is a constant number of hydrogen bonds between the taa1/taa3 group .
+ significant fluctuations in the cys1/cys3 and the cys2/cys4 group after 600 ps can be observed which may be responsible for the opening of the taa1/taa3 group on a timescale up to 3 ns . finally on longer timescales than 3 ns the structure unfolds into a fully extended strand in the 500 k simulations as it can be seen by the large rmsd of all groups .
it is remarkable that the 300 k simulations do not reach the fully unfolded structure as the inset of fig .
[ fig2 ] indicates .
hence it becomes clear that the usage of biasing methods at these lower temperatures is required . +
a more detailed analysis of the unfolding pathway can be performed by the investigation of the essential dynamics @xcite . to study the concerted motion of the molecule and to indicate the main structural transitions we analyzed the essential eigenvectors of the system .
most of the motion at 500 k which is about 66% is expressed in the first two eigenvectors as fig .
[ fig3 ] indicates .
nearly 57% of the overall motion is incorporated in eigenvector 1 .
the essential subspace is formed by the first eight eigenvectors with roughly 88% . the relative positional fluctuations again indicate that eigenvector 1 is dominating the overall motion with significant contributions from eigenvectors 2 and 3 .
the presence of all eigenvectors larger than 8 is negligible .
it can be concluded that this represents an unfolding pathway which can be effectively described by a small number of eigenvectors .
+ the motion of the first two eigenvectors at 500 k is presented in fig .
eigenvector 1 mainly describes the opening of the end - to - end distance by a stretching mode .
eigenvector 2 can be identified as the initial opening of the i - motif by the relaxation into a planar structure .
+ to compare the unfolding pathways of 500 k to lower temperatures , we performed a 50 ns metadynamics simulation at 300 k. this allows to overcome energetic barriers and to identify the main unfolding pathways which can be compared to the high temperature simulations by an analysis of the essential dynamics .
+ for this we investigated the eigenvectors for the biased metadynamics run at 300 k , the unbiased 500 k and 300 k simulations .
the projection is performed along the unbiased 500 k subset of eigenvectors 1 and 2 on a timescale of 10 ns . + for the biased 300 k simulations and the unbiased 500 k simulations . ,
scaledwidth=50.0% ] fig .
[ fig5 ] illustrates the importance of elaborated temperatures respectively the usage of biasing potentials to discover the unfolding pathway .
it can be seen that the biased simulations at 300 k resembles the high temperature trajectory in good agreement although it is limited to a smaller region .
this is also true for the unbiased 300 k simulation which is restricted to a very small subspace due to occurring trapping in a specific configuration whereas the main unfolding pathway is identical .
the results indicate that the unfolding mechanism is remarkably accelerated by elaborated temperatures or biasing methods and obviously resembles the real low temperature unfolding pathway .
+ the results above are additionally validated by the calculation of the squared inner product @xmath44 ( eqn .
[ eq : solp ] ) for the complete biased 300 k and the unbiased 500 k simulations and the eigenvector inner product @xmath45 which is presented on the right side of fig .
the squared inner product for the first three eigenvectors is @xmath46 with a square root of @xmath47 which indicates a nearly identical subspace of the first three eigenvectors at different temperatures .
these values are supported by the inner product matrix @xmath45 .
eigenvector 1 has an inner product of @xmath48 indicating a remarkable identity of both trajectories .
this is also true for eigenvectors 2 and 3 with values of @xmath49 and @xmath50 .
hence the overall motion of the first three eigenvectors which describe 73 % of the concerted motion ( fig .
[ fig4 ] ) is directly comparable at both temperatures .
this proves the assumption that the usage of elaborated temperatures results in unfolding pathways that are comparable to those derived at lower temperatures .
+ the results for all unbiased 500 k simulations and the complete biased 300 k simulation trajectory are presented in fig .
it is obvious that similar regions are explored at both temperatures and the overall displacement at both temperatures is comparable .
several conformations can be identified along the trajectories starting with the initial i - motif . for a detailed analysis of the stability of the presented conformations and the unfolding pathway , the underlying free energy landscapes have to be calculated .
this allows to decide whether a conformation is stable or represents a global minimum which can be directly related to the detected unfolding path .
the free energy landscapes at 300 k and 500 k are presented in fig .
[ fig7 ] . + it is obvious that the most stable conformation is given by a stretched structure at 500 k. this is in contrast to 300 k where it can be seen that two stable hairpin conformations ( b ) and ( c ) represent the lowest free energy conformations .
it can be assumed that the local ordering of water molecules is mainly responsible for this effect due to the fact that it is largely dominated by electrostatic interactions of the dna . at larger temperatures ,
the dominance of the coulomb interactions in comparison to thermal energy becomes smaller such that additional solvent and molecular configurations are accessible .
finally this results in a change of the entropy of these configurations such that the free energy minima are adjusted by entropic contributions .
although the minima and the barriers differ in their height and their relative position , the overall shape of the landscapes is similar .
entropic contributions as discussed above may be responsible for these slight deviations .
the good agreement becomes also obvious by studying the unfolding pathways and the corresponding free energies .
thus it can be concluded that comparable unfolding pathways can be observed as it was also discussed above .
however , the realization of the unfolding pathway in an unbiased simulation at lower temperatures is dependent on the energy barriers that are occurring on the pathway . as fig .
[ fig7 ] indicates , the general movement of the unfolding process is comparable at both temperatures .
+ the hairpin conformations ( b ) and ( c ) at 300 k belong to free energy minima and are the reason for the stagnating root - mean square deviation presented in fig .
hence the overcome of the energetic barriers for these structures is cumbersome in a 300 k unbiased simulation .
thus it can be concluded that the unfolding pathway at lower temperatures is mainly realized by a transition into conformations ( b ) and ( c ) .
in contrast to 300 k , the free energy landscape at 500 k indicates these conformations as local minima with low barriers such that the transition into a fully extended strand representing the global energy minimum can be realized .
+ starting with the i - motif ( a ) the initial opening of one end with the groups cys3-t is sufficient for the transition in the global minimum ( b ) at 300 k. overcoming this stable configuration results in two possible pathways .
the first possibility is the transition into the neighbored minimum which is given by a planar hairpin structure ( c ) .
the other pathway follows the opening of the strand with groups cys1 and taa1 into a planar s - shaped structure ( d ) .
the final transition is then realized by an unfolding into the fully extended strand ( e ) .
however , the last pathway is better defined for 300 k due to the fact that the completely unfolded region is energetically more favorable for the 500 k simulations as the free energy landscapes indicate .
however all configurations have been clearly visited in the 500 k simulations . +
a further pathway is the unfolding into a twisted structure ( g ) where the strand with groups cys1-t follows a rotational motion .
after this move , the ends of the strand relax into structure ( f ) . by a second process
, the twisted configuration transforms into conformation ( d ) which is also present in the first unfolding pathway at 300 k. the final transition into the stretched structure ( e ) is identical to the first unfolding pathway . at 500
k this pathway visits conformations ( g ) until the configuration fully unfolds on different pathways as discussed above .
we have simulated the unfolding of the dna i - motif via molecular dynamics simulations in explicit solvent at 500 k and at 300 k. our results indicate that the initial structure of the i - motif vanishes on a timescale of a few nanoseconds .
thus our results are in agreement to experimental observations which indicated that an unprotonated i - motif is not stable at room temperature @xcite .
+ we have shown that unfolding pathways can be investigated either by the usage of elaborated temperatures as well as biasing potentials .
the occurring trajectories are similar as a detailed analysis of the essential dynamics indicated .
hence we have presented a method to prove the transformation of high temperature unfolding trajectories onto lower temperatures . due to this similarity
we can identify two main unfolding pathways for the dna i - motif at both temperatures .
the most relevant one at 300 k visits stable hairpin configurations from which it unfolds only due to thermal activation or the presence of biasing potentials into an extended structure in a reasonable simulation time .
the other one includes torsional motion and is energetically less favorable . + the calculated free energy landscapes at 500 k and 300 k are similar in their global shape which leads to comparable unfolding pathways .
hence it can be concluded that this property is the main reason for the success of high temperature simulations .
finally our results indicate that the analysis of the unfolding pathways is only reasonable in combination with the calculated free energy landscapes .
this becomes specifically important by comparing the global energetic minima which may result in only partly visited unfolding pathways . hence it can be concluded that at lower temperatures the i - motif mainly unfolds into the discussed hairpin configurations .
the fully extended strand represents the global energy minimum only at higher temperatures .
finally it is obvious that the detailed unfolding pathway is better defined for 300 k although the main configurations are visited at both temperatures . + as it has been discussed , the determination of the unfolding pathway of the i - motif is of prior importance regarding the applicability in nanomachines @xcite .
hence our results allow to improve the applicability of i - motifs as functional materials in nanotechnology due to the explicit knowledge of the unfolding pathway and its stable conformations .
finally the investigation of the unfolding properties may also lead to a more detailed knowledge about the molecular mechanisms in the human cell .
this is in particular interesting due to the fact that the knowledge of the function of the i - motif is still missing .
we thank the deutsche forschungsgemeinschaft ( dfg ) through the transregional collaborative research center trr 61 for financial funding .
99 daggett v. and levitt m. , _ j. mol .
biol . _ * 323 * , 600 ( 1993 ) day r. , bennion b. j. , ham s. and daggett v. , _ j. mol .
biol . _ * 322 * , 189 ( 2002 ) daggett v. and fersht a. , _ rev .
cell biol . _
* 4 * , 497 ( 2003 ) beck d. a. c. and daggett v. , _ methods _ * 34 * , 112 ( 2004 ) dobson c. m. , sali a. and karplus m. , _ angew .
ed . _ * 37 * , 868 ( 1998 ) wang t. and wade r. c. , _
theory comput .
_ * 3 * , 1476 ( 2007 ) gehring k. , leroy j .-
l . and gueron m. , _ nature _ * 363 * , 561 ( 1993 ) gueron m. and leroy j .-
_ * 10 * , 326 ( 2000 ) lane a. n. , chaires j. b. , gray r. d. trent j. o. , _ nucleic acids res . _ * 36 * , 5482 ( 2008 ) liu h. and liu d. , _ chem .
* 19 * , 2625 ( 2009 ) phan a. t. and mergny j .-
_ nucleic acids res .
_ * 30 * , 4618 ( 2002 ) zhao y. , zeng z .- x . ,
kan z .- y . ,
hao y .- h . and tan z. , _ chembiochem _ * 6 * , 1957 ( 2005 ) catasti p. , chen x. , deaven l. l. , moyzis r. k. , bradbury e. m. and gupta g. , _ j. mol
. biol . _ * 272 * , 369 ( 1997 ) hurley l. h. , _ biochem .
trans . _ * 29 * , 692 ( 2001 ) liu d. and balasubramian s. , _ angew .
_ * 42 * , 5734 ( 2003 ) mao y. , liu d. , wang s. , luo s. , wang w. , yang y. , ouyang q. and jiang l. , _ nucleic acids res . _ * 35 * , 33 ( 2007 ) modi s. , swetha m. g. , goswami d. , gupta g. d. , mayor s. and krishnan y. , _ nat . nanotechnol . _
* 4 * , 325 ( 2009 ) pu f. , wang c. y. , hu d. , huang z. z. , ren j. s. , wang s. and qu x. g. , _ mol .
_ 6 _ , 1928 ( 2010 ) peng y. h. , wang x. h. , xiao y. , feng l. y. , zhao c. , ren j. s. and qu x. g. , _ j. am .
_ * 131 * , 13813 ( 2009 ) xia h. , hou y. , ngai y. and zhang g. , _ j. phys
b _ _ 114 _ , 775 ( 2010 ) laio a. and parrinello m. , _ proc . natl .
usa _ * 99 * , 12562 ( 2002 ) laio a. and gervasio f. l. , _ rep .
. phys . _ * 71 * , 126601 ( 2008 ) ensing b. , de vivo m. , liu z. , moore p. and klein m. l. , _ acc . chem .
* 39 * , 73 ( 2005 ) ensing b. , laio a. , parrinello m. and klein m. l. , _ j. phys .
b _ * 109 * , 6676 ( 2006 ) smiatek j. and heuer a. , _ j. comp . chem . _ , to be published kumar s. , rosenberg j. m. , bouzida d. , swenden r. and kollman p. a. , _ j. comput
* 16 * , 1339 ( 1992 ) amadei a. , linssen a. b. and berendsen h. j. c. , _ proteins _ * 17 * , 412 ( 1993 ) spiwok v. , lipova p. and kralova b. , _ j. phys . chem .
b _ * 111 * , 3073 ( 2007 ) spiwok v. , kralova b. and tvaroska i. , _ j. mol
. model _ * 14 * , 995 ( 2008 ) de groot b. l. , van aalten d. m. f. , amadei a. and berendsen h. j. c. , _ biophys . j. _ * 71 * , 1707 ( 1996 ) van aalten d. m. f. , de groot b. l. , findlay j. b. c. , berendsen h. j. c. and amadei a. , _ j. comp . chem . _
* 18 * , 169 ( 1997 ) amadei a. , ceruso m. a. and di nola a. , _ proteins _ * 36 * , 419 ( 1999 ) de groot b. l. , daura x. , mark a. e. abd grubmller h. , _ j. mol .
biol . _ * 309 * , 299 ( 2001 ) hess b. , _ phys .
e _ * 65 * , 031910 ( 2002 ) hess b. , kutzner c. , van der spoel d. and lindahl e. , _
theory comput . _
* 4 * , 435 ( 2008 ) phan t. a. , gueron m. and leroy j .-
l . , _ j. mol .
* 299 * , 123 ( 2000 ) sorin e. j. , pande v. s. , _ biophys .
j. _ * 88 * , 2472 - 2493 ( 2005 ) smiatek j. , liu d. and heuer a. , _ to be published _ ( 2011 ) bonomi m. , branduardi d. , bussi g. , camilloni c. , provasi d. , raiteri p. , donadio d. , marinelli f. , pietrucci f. , broglia r. a. and parrinello m. , _ comp .
* 180 * , 1961 ( 2009 ) laio a. , rodriguez - fortea a. , gervasio f. l. , ceccarelli m. and parrinello m. , _ j. phys .
b _ * 109 * , 6714 ( 2005 ) | we present the results of high temperature 500 k molecular dynamics simulations of the dna i - motif .
the essential dynamics and the main unfolding pathways are compared to a biased metadynamics simulation at 300 k. our results indicate a remarkable agreement of the concerted motion at both temperatures .
the transition can be described by a few number of eigenvectors indicating a simple unfolding mechanism .
two main mechanisms for the unfolding pathway at 500 k can be detected which are in good agreement to the results of the biased simulation at 300 k. |
the last century was marked by the birth of two major scientific and engineering disciplines : silicon - based computing and the theory and technology of genetic data analysis .
the research field very likely to dominate the area of scientific computing in the foreseeable future is the merger of these two disciplines , leading to unprecedented possibilities for applications in varied areas of engineering and science .
the first steps toward this goal were made in 1994 , when leonard adleman @xcite solved a quite unremarkable computational problem , an instance of the directed travelling salesmen problem on a graph with seven nodes , with an exceptional method .
the technique used for solving the problem was a new technological paradigm , termed dna computing .
dna computing introduced the possibility of using genetic data to tackle computationally hard classes of problems that are otherwise impossible to solve using traditional computing methods .
the way in which dna computers make it possible to achieve this goal is through massive parallelism of operation on nano - scale , low - power , molecular hardware and software systems .
one of the major obstacles to efficient dna computing , and more generally dna storage @xcite and signal processing @xcite , is the very low reliability of single - stranded dna sequence operations .
dna computing experiments require the creation of a controlled environment that allows for a set of single - stranded dna codewords to bind ( hybridize ) with their complements in an appropriate fashion .
if the codewords are not carefully chosen , unwanted , or non - selective , hybridization may occur .
even more detrimental is the fact that a single - stranded dna sequence may fold back onto itself , forming a secondary structure which completely inhibits the sequence from participating in the computational process .
secondary structure formation is also a major bottleneck in dna storage systems .
for example , it was reported in @xcite that @xmath0 of read - out attempts in a dna storage system failed due to the formation of special secondary structures called _ hairpins _ in the single - stranded dna molecules used to store information .
so far , the focus of coding for dna computing was exclusively directed towards constructing large sets of dna codewords with fixed base frequencies ( constant gc - content ) and prescribed hamming / reverse - complement hamming distance .
such sets of codewords are expected to result in very rare hybridization error events . as an example
, it was shown in @xcite that there exist @xmath1 codewords of length @xmath2 with minimum hamming distance @xmath3 and with exactly @xmath4 /@xmath5 bases . at the same time , the wisconsin dna group , led by d. shoemaker , reported that through extensive computer search and experimental testing , only @xmath6 sequences of length @xmath2 at hamming distance at least @xmath7 were found to be free of secondary structure at temperatures of @xmath8 . since at lower ambient temperatures the probability of secondary structure formation is even higher , it is clear that the effective number of codewords useful for dna computing is extremely small . in this paper
, we investigate properties of dna sequences that may lead to undesirable folding . our approach is based on analysis of a well - known algorithm for approximately determining dna secondary structure , called nussinov s method .
this analysis allows us to extract some design criteria that yield dna sequences that are unlikely to fold undesirably .
these criteria reduce to the requirement that the first few shifts of a dna codeword have the property that they do not contain watson - crick complementary matchings with the original sequence .
we consider the enumeration of sequences having the shift property and provide some simple construction strategies which meet the required restrictions . to the best of our knowledge , this is the first attempt in the literature aimed at providing a rigorous setting that links dna folding properties to constraints on the primary structure of the sequences .
dna of higher species consists of two complementary chains twisted around each other to form a double helix .
each chain is a linear sequence of nucleotides , or _
bases_ two _ purines _ , adenine ( ) and guanine ( ) , and two _ pyrimidines _ , thymine ( ) and cytosine ( ) . the purine bases and pyrimindine bases are _ watson - crick ( wc ) complements _ of each other , in the sense that @xmath9 more specifically , in a double helix , the base pairs with by means of two hydrogen bonds , while pairs with by means of three hydrogen bonds ( i.e. the strength of the former bond is weaker than the strength of the latter ) . for dna computing purposes ,
one is only concerned with single - stranded ( henceforth , _ oligonucleotide _ ) dna sequences .
oligonucleotide dna sequences are formed by heating dna double helices to denaturation temperatures , at which they break down into single strands .
if the temperature is subsequently reduced , oligonucleotide strands with large regions of sequence complementarity can bind back together in a process called _ hybridization_. hybridization is assumed to occur only between complementary base pairs , and lies at the core of dna computing . as a first approximation
, oligonucleotide dna sequences can be simply viewed as words over a four - letter alphabet @xmath10 , with a prescribed set of complex properties .
the generic notation for such sequences will be @xmath11 , with @xmath12 indicating the length of the sequences .
the wc complement @xmath13 of a dna sequence is defined as @xmath14 , @xmath15 being the wc complement of @xmath15 as given by ( [ bar_def ] ) .
the _ secondary structure _ of a dna codeword @xmath16 is a set , @xmath17 , of disjoint pairings between complementary bases @xmath18 with @xmath19 .
a secondary structure is formed by a chemically active oligonucleotide sequence folding back onto itself due to _ self - hybridization _ , _
i.e. _ , hybridization between complementary base pairs belonging to the same sequence . as a consequence of the bending ,
elaborate spatial structures are formed , the most important components of which are loops ( including branching , internal , hairpin and bulge loops ) , stem helical regions , as well as unstructured single strands .
figure [ sec - str ] illustrates these concepts for an rna strand .
it was shown experimentally that the most important factors influencing the secondary structure of a dna sequence are the number of base pairs in stem regions , the number of base pairs in a hairpin loop region as well as the number of unpaired bases .
determining the exact pairings in a secondary structure of a dna sequence is a complicated task , as we shall try to explain briefly . for a system of interacting entities ,
one measure commonly used for assessing the system s property is the free energy .
the stability and form of a secondary configuration is usually governed by this energy , the general rule - of - thumb being that a secondary structure minimizes the free energy associated with a dna sequence .
the free energy of a secondary structure is determined by the energy of its constituent pairings .
now , the energy of a pairing depends on the bases involved in the pairing as well as all bases adjacent to it . adding complication is the fact that in the presence of other neighboring pairings , these energies change according to some nontrivial rules .
nevertheless , some simple dynamic programming techniques can be used to _
approximately _ determine the secondary structure of a dna sequence .
such approximations usually have the correct form in @xmath20 of the cases considered . among these techniques , _
nussinov s folding algorithm _ is the most widely used scheme @xcite .
nussinov s algorithm is based on the assumption that in a dna sequence @xmath21 , the energy between a pair of bases , @xmath22 , is independent of all other pairs .
for simplicity of exposition , we shall assume that @xmath23 if @xmath24 , @xmath25 if @xmath26 , and @xmath27 otherwise and @xmath28 . ] .
let @xmath29 denote the minimum free energy of the subsequence @xmath30 .
the independence assumption allows us to compute the minimum free energy of the sequence @xmath31 through the recursion @xmath32 where @xmath33 for @xmath34 and @xmath35 for @xmath36 .
the value of @xmath37 is the minimum free energy of a secondary structure of @xmath38 .
note that @xmath39 .
a very large negative value for the free energy @xmath37 of a sequence is a good indicator of the presence of stacked base pairs and loops , _
i.e. _ , a secondary structure , in the physical dna sequence .
nussinov s algorithm can be described in terms of free - energy tables , two of which are shown below .
we first describe how such a table is filled out , after which we will point out some important properties of such tables .
.free - energy table for the sequence [ cols="^,^,^,^,^,^,^,^,^,^ " , ] in a free - energy table , the entry at position @xmath40 ( the top left position being ( 1,1 ) ) , contains the value of @xmath29 .
the table is filled out by initializing the entries on the main diagonal and on the first lower sub - diagonal of the matrix to zero , and calculating the energy levels according to the recursion in .
the calculations proceed successively through the upper diagonals : entries at positions @xmath41 are calculated first , followed by entries at positions @xmath42 , and so on .
note that the entry at @xmath43 , depends on @xmath44 and the entries at @xmath45 , @xmath46 , @xmath47 , @xmath48 , and @xmath49 .
the minimum - energy secondary structure itself can be found by the _ backtracking algorithm _
@xcite which retraces the steps of nussinov s algorithm .
the secondary structures for the sequences in tables [ table1 ] and [ table2 ] , shown in figures [ sec - str1 ] and [ sec - str2 ] , have been found using the vienna rna / dna secondary structure package @xcite , which is based on the nussinov algorithm , but which uses more accurate values for the parameters @xmath22 , as well as more sophisticated prediction methods for base pairing probabilities .
tables [ table1 ] and [ table2 ] show that the minimum free energy for the sequence is @xmath50 , while that for the sequence is @xmath51 .
, the sequence is deemed to have no secondary structure ; this is due to the fact that the one possible complementary base pairing , namely that of and , forms too weak a bond for the resultant structure to be stable .
] this fact alone indicates that the number of paired bases in the first sequence ought to be larger than in the second one , and hence the former is more likely to have a secondary structure than the latter .
more generally , one can observe the following characteristics of free - energy tables : if the first upper diagonal contains a large number of @xmath51 or @xmath28 entries , then some of these entries `` percolate '' through to the second upper diagonal , where they get possibly increased by @xmath51 or @xmath28 if complementary base pairs are present at positions @xmath52 and @xmath53 , @xmath54 in the dna sequence .
the values on the second diagonal , in turn , percolate through to the third diagonal , and so on .
hence , the free energy of the dna sequence depends strongly on the number of non - zero values present on the first diagonal .
this phenomenon was also observed experimentally in @xcite , where the free energy was modelled by a function of the form @xmath55 with @xmath56 denoting a correction factor which depends on the number of @xmath57 and @xmath58 bases in the sequence * c*. the stability of a secondary structure , as well as its melting properties can be directly inferred from ( [ app - energy ] ) .
note that under the assumption that @xmath59 for all pairings , the absolute value of the sum in ( [ app - energy ] ) is simply the total number of pairings of complementary bases between the dna codeword * c * and the codeword shifted one position to the right or equivalently , the sum of the entries in the first upper diagonal of the free - energy table .
these observations imply that a more accurate model for the free energy should be of the form @xmath60 for a correction factor @xmath56 and some non - zero weighting factors @xmath61 .
furthermore , the same observation implies that from the stand - point of designing dna codewords without secondary structure , it is desirable to have codewords for which the respective sums of the elements on the first several diagonals are either all zero or of some very small absolute value .
this requirement can be rephrased in terms of requiring a dna sequence to satisfy a _
shift property _ , in which a sequence and its first few shifts have few or no complementary base pairs at the same positions .
in this section , we consider the enumeration and construction of dna sequences satisfying certain shift properties , which we shall define rigorously . it is clear that a dna sequence is counted by @xmath81 iff it contains no pair of complementary bases
. such a sequence must be over one of the alphabets @xmath84 , @xmath85 , @xmath86 and @xmath87 .
there are @xmath88 such sequences , since there are @xmath89 sequences over each of these alphabets , of which @xmath90 , @xmath91 , @xmath92 and @xmath93 are each counted twice .
let @xmath96 denote the set of all sequences @xmath75 of length @xmath12 for which @xmath78 , @xmath97 .
thus , @xmath98 .
note that for any @xmath99 , @xmath100}$ ] can not contain a complementary pair of bases , and hence can not contain three distinct bases .
let @xmath101 denote the set of sequences @xmath102 such that @xmath103 , and let @xmath104 .
we thus have @xmath105 . each sequence in @xmath101
is obtained from some sequence @xmath106 by appending @xmath107 bases , @xmath108 , all equal to @xmath109 .
hence , @xmath110 , and therefore , @xmath111 .
now , observe that each sequence @xmath112 is obtained by appending a single base , @xmath113 , to some sequence @xmath114 .
if @xmath115 is in fact in @xmath116 , then there are three choices for @xmath117 . otherwise ,
if @xmath118 , there are only two possible choices for @xmath117 .
hence , @xmath119 this proves the claimed result .
it can be shown that for @xmath122 , the polynomial @xmath123 in the denominator of @xmath124 has a real root , @xmath125 , in the interval ( 2,3 ) , and @xmath107 other roots within the unit circle .
it follows that @xmath126 for some constant @xmath127 .
it is easily seen that @xmath125 decreases as @xmath128 increases , and that @xmath129 .
let @xmath132 be the set of length-@xmath12 dna sequences @xmath75 such that @xmath130 .
a sequence @xmath133 is in @xmath132 iff the set @xmath134 has cardinality @xmath135 .
so , to construct such a sequence , we first arbitrarily pick a @xmath136 and an @xmath137 , @xmath138 , which can be done in @xmath139 ways .
the rest of @xmath75 is constructed recursively : for @xmath140 , set @xmath141 if @xmath142 , and pick a @xmath143 if @xmath144 .
thus , there are 3 choices for each @xmath140 , @xmath144 , and hence a total of @xmath131 sequences @xmath75 in @xmath132 . a dna code in which all codewords have the same gc - content , @xmath147 ,
is called a _ constant gc - content code_. the constant gc - content constraint is introduced in order to achieve parallelized operations on dna sequences , by assuring similar thermodynamic characteristics of all codewords .
the gc - content usually needs to be in the range of @xmath148 of the length of the code .
the problem of constructing dna codewords obeying some form of a shift constraint can be reduced to a binary code design problem by mapping the dna alphabet onto the set of length - two binary words as follows : @xmath152 let * q * be a dna sequence .
the sequence @xmath153(*q * ) obtained by applying coordinatewise the mapping given in to * q * will be called the _ binary image _ of * q*. if @xmath154 , then the subsequence @xmath155 will be referred to as the _ even subsequence _ of @xmath156 , and @xmath157 will be called the _ odd subsequence _ of @xmath153(*q * ) . the wc distance , @xmath158 , between two dna words @xmath159 can be expressed in terms of the even and odd subsequences , as stated in the lemma below . for notational
ease , given binary words @xmath160 and @xmath161 , we define @xmath162 , the sum being taken modulo-2 , and @xmath163 .
let * p * and * r * be two words of length @xmath12 over the alphabet @xmath164 , and define @xmath165 , @xmath166 . using @xmath167 to denote the complement of a binary sequence @xmath168 , we can express the wc distance between @xmath169 and @xmath170 as @xmath171 where @xmath172 denotes hamming weight .
consequently , if for some length-@xmath12 dna sequence @xmath173 , we take @xmath174}$ ] and @xmath175}$ ] , then @xmath173 satisfies the @xmath52-th shift constraint @xmath176 iff @xmath177 [ wc - binary ] in a companion paper @xcite , we described a sample of construction methods for dna codes which reduce undesired hybridization and alow for fully parallel dna system operation . among the constraints identified for this problem
are the base runlength constraint , the constant gc - content constraint , and the hamming and reverse - complement hamming distance constraint .
we will show next that it is straightforward to incorporate these hybridization constraints into a scheme which also allows for constructing sequences with reduced probability of secondary structure formation .
the idea is based on the use of the non - zero codewords of cyclic simplex codes ( * ? ? ?
* chapter 8) or subsets thereof . recall that a cyclic simplex code of dimension @xmath135 is a simplex code of length @xmath178 composed of the all - zeros codeword and the @xmath12 distinct cyclic shifts of any non - zero codeword .
let @xmath179 be a dna code obtained by choosing the set of non - zero codewords of a cyclic simplex code of length @xmath180 for both the even and odd binary code component .
such a code contains @xmath181 codewords with the property that for all @xmath182 and @xmath183 , @xmath184 [ simplex_theorem ] it is straightforward to see that for any @xmath185 , if we let @xmath174}$ ] , @xmath186}$ ] , then @xmath187 and @xmath188 , defined as in lemma [ wc - binary ] , are just truncations of codewords from the simplex code .
since the simplex code is a constant - weight code , with minimum distance @xmath189 , each pair of codewords intersects in exactly @xmath189 positions .
this implies that there exist exactly @xmath189 positions for which one given codeword contains all zeros , and the other codeword contains all ones .
these are the positions that are counted in @xmath190 , which proves the claimed result .
consider the previous construction for @xmath191 , and a generating codeword @xmath192 .
there are @xmath193 dna codewords of length @xmath194 obtained based on the outlined method .
these codewords have minimum hamming distance equal to four , and they also have constant @xmath195 content @xmath196 . a selected subset of codewords from this code is listed below .
@xmath197 the last two codewords consist of the bases @xmath5 and @xmath5 only , and clearly satisfy the shift property with @xmath198 for all @xmath199 . on the other hand , for the first three codewords
one has @xmath200 , while for the next three codewords it holds that @xmath201 ( meeting the upper bound in the theorem ) . due to the cyclic nature of the generating code
, one can easily generate the nussinov folding table for all the codewords @xcite .
such an evaluation , as well as the use of the program package vienna , show that none of the @xmath193 codewords exhibits a secondary structure .
the largest known dna codes with the parameters specified above consists of @xmath202 codewords @xcite .
this code is generated by a simulated annealing process which does not allow for simple secondary structure testing .
adleman , `` molecular computation of solutions to combinatorial problems , '' _ science _ ,
266 , pp.10211024 , nov . 1994 .
k. breslauer , r. frank , h. blocker , and l. marky , `` predicting dna duplex stability from the base sequence , '' _ proceedings of the national academy of science _
, usa 83 , pp . 37463750 , 1986 .
p. gaborit , and h. king , `` linear constructions for dna codes , '' preprint .
goulden and d.m .
jackson , _ combinatorial enumeration _ ,
dover , 2004 .
hall , lecture notes on error - control coding , available online at ` http://www.mth.msu.edu/\simjhall/ ` .
macwilliams , and n.j.a .
sloane , _ the theory of error correcting codes _ , north - holland , 1977 .
m. mansuripur , p.k .
khulbe , s.m .
kuebler , j.w .
perry , m.s.giridhar and n. peyghambarian , `` information storage and retrieval using macromolecules as storage media , '' _ university of arizona technical report _ , 2003 .
o. milenkovic and n. kashyap , `` new constructions of codes for dna computing , '' accepted for presentation at wcc 2005 , bergen , norway .
s. mneimneh , `` computational biology lecture 20 : rna secondary structures , '' available online at ` engr.smu.edu/\simsaad/courses/ cse8354/lectures / lecture20.pdf ` .
r. nussinov , g. pieczenik , j.r .
griggs and d.j .
kleitman , `` algorithms for loop matchings , ''
_ siam j. appl .
_ , vol .
35 , no . 1
, pp . 6882 , 1978 . s. tsaftaris , a. katsaggelos , t. pappas and e. papoutsakis , `` dna computing from a signal processing viewpoint , '' _ ieee signal processing magazine _ , pp .
100106 , sept .
2004 . the vienna rna secondary structure package , ` http://rna.tbi.univie.ac.at/cgi-bin/rnafold.cgi ` | in this paper , we consider the problem of designing codewords for dna storage systems and dna computers that are unlikely to fold back onto themselves to form undesirable secondary structures .
secondary structure formation causes a dna codeword to become less active chemically , thus rendering it useless for the purpose of dna computing .
it also defeats the read - back mechanism in a dna storage system , so that information stored in such a folded dna codeword can not be retrieved .
based on some simple properties of a dynamic - programming algorithm , known as nussinov s method , which is an effective predictor of secondary structure given the sequence of bases in a dna codeword , we identify some design criteria that reduce the possibility of secondary structure formation in a codeword .
these design criteria can be formulated in terms of the requirement that the watson - crick distance between a dna codeword and a number of its shifts be larger than a given threshold .
this paper addresses both the issue of enumerating dna sequences with such properties and the problem of practical dna code construction . |
with results from the sample experiment available @xcite , there is now great interest in parity - violating electron - deuteron scattering . the sample collaboration measures at bates the longitudinal asymmetry in polarized quasi - elastic electron scattering on the proton @xcite and deuteron @xcite .
the asymmetries are sensitive to the nucleon s strange magnetic and axial form factors .
if other effects are under control , the two measurements can be used to determine the values of the form factors at @xmath6 gev@xmath2 . neglecting two - nucleon current effects , preliminary results are in disagreement with theoretical estimates , in particular for the axial contribution @xcite .
the purpose of the work presented here is to compute a theoretical value for the deuteron asymmetry including two - nucleon currents .
neutral charge and current one - nucleon operators are well known ( see , for example , ref .
the contribution of such operators to the asymmetry was computed at various momentum transfers in ref .
several theoretical issues were studied in detail . for the kinematical region relevant to sample
, final - state interactions were found to be important .
it was also found that ( except for very low momentum transfers ) the asymmetry in the vicinity of the quasi - elastic peak is fairly independent of the choice of two - nucleon ( @xmath7@xmath7 ) potential .
two - nucleon charge and current operators have also been studied to some extent .
heavy - meson exchange contributions were considered in ref .
they were shown to be unimportant in a calculation of the asymmetry that neglected final - state interactions . in ref .
@xcite , an impulse approximation modified to incorporate gauge invariance was employed .
the effects of parity - violating @xmath7@xmath7 interactions on the deuteron wave function were found to be small .
pion - exchange currents were included in the computation of the asymmetry , but only in electromagnetic sector . in the present work
we extend these earlier calculations .
the leading currents are calculated in chiral perturbation theory ( chpt ) @xcite supplemented by a successful phenomenological model @xcite .
the one - nucleon currents considered here include phenomenological form factors and have the same form as those in ref .
we also consider pion - exchange currents in both electromagnetic and weak contributions , as well as effects from heavier mesons , evaluated using the riska prescription @xcite .
the asymmetry is calculated with deuteron and final - state wave functions obtained from a realistic potential , the argonne @xmath4 model @xcite .
we present results for the kinematical region of interest to sample .
( our calculation can be repeated at other momentum transfers , such as those of the planned sample - lite experiment @xcite and jlab s g0 experiment @xcite . )
we verify that when two - nucleon effects are turned off , our results are in agreement with those of ref .
we study the effects of two - body currents on the asymmetry , both near the quasi - elastic peak where one - body processes should dominate , and away from the quasi - elastic peak where these two - body currents could be important .
we are able to confirm that most of the two - nucleon contributions to the asymmetry are due to currents of pion range .
our results show that near the quasi - elastic peak two - body currents give a small contribution to the asymmetry . away from the peak ,
they become more important , and can increase the magnitude of the asymmetry by as much as 3% .
we find that the contribution to the asymmetry associated with the electromagnetic - axial current interference response function is about 20% .
the overall effect of two - nucleon currents on the data of the sample experiment is indeed small but not negligible , and is being incorporated in the data analysis @xcite .
a summary of the relevant formulas for the calculation of the asymmetry is given in sec .
[ sec : asy ] .
section [ sec : emop ] contains the one- and two - body electromagnetic currents , and sec .
[ sec : weop ] contains the one- and two - body neutral weak currents .
the connection between our phenomenological approach and chpt can be found in sec . [
sec : wechi ] , while the numerical computation of the asymmetry is described in sec .
[ sec : calc ] .
results and conclusions are presented in sec .
[ sec : res ] .
parity - violating electron - nucleus scattering results from interference of amplitudes associated with photon and @xmath0 exchanges , shown in fig . [ fig : ampl ] . the initial and final electron ( nucleus ) four - momenta
are labelled by @xmath8 and @xmath9 ( @xmath10 and @xmath11 ) , respectively , while the four - momentum transfer @xmath12 is defined as @xmath13 .
the amplitudes for the processes in fig .
[ fig : ampl ] are then given by @xcite @xmath14 @xmath15 @xmath16 where @xmath17 and @xmath18 are the fine - structure constant and fermi constant for muon decay , respectively , @xmath19 and @xmath20 are the standard model values for the neutral - current couplings to the electron given in terms of the weinberg angle @xmath21 , @xmath22 and @xmath23 are the initial and final electron spinors , and @xmath24 and @xmath25 denote matrix elements of the electromagnetic and weak neutral currents , i.e. @xmath26 and similarly for @xmath25 . here
@xmath27 and @xmath28 are the initial and final nuclear states .
note that in the amplitude @xmath29 the @xmath30 dependence of the @xmath0 propagator has been ignored , since here we restricted ourselves to @xmath31 .
.25 cm the parity - violating asymmetry in the quasi - elastic regime is given by @xmath32 where @xmath33 is the inclusive cross section for scattering of an incident electron with helicity @xmath34 .
it is easily seen that , to leading order , @xmath35 standard manipulations then lead to the following expression for the asymmetry in the extreme relativistic limit for the electron @xcite @xmath36 where the @xmath37 s are defined in terms of electron kinematical variables , @xmath38 @xmath39 being the electron scattering angle in the laboratory , while the @xmath40 s are the nuclear electro - weak response functions , which depend on @xmath41 and @xmath42 , to be defined below . to this end , it is first convenient to separate the weak current @xmath43 into its vector @xmath44 and axial - vector @xmath45 components , and to write correspondingly @xmath46 the response functions can then be expressed as @xmath47 \ , \label{eq : rl}\ ] ] @xmath48 \ , \label{eq : rt}\ ] ] @xmath49 \ , \label{eq : rtp}\ ] ] where @xmath50 is the mass of the target ( assumed at rest in the laboratory ) , @xmath51 is the energy of the final nuclear state ( in general , a scattering state ) , and in eqs .
( [ eq : rl ] ) and ( [ eq : rt ] ) the superscript @xmath52 is either @xmath53 or @xmath54 . note that there is a sum over the final states and an average over the initial spin projection states of the target , as implied by the notation @xmath55 . in the expressions above for the @xmath40 s
, it has been assumed that the three - momentum transfer @xmath56 is along the @xmath57-axis , which defines the spin quantization axis for the nuclear states . the model for the nuclear electro - weak currents
is discussed in the next sections , while the calculation of the deuteron response functions is described in sec .
[ sec : calc ] .
the nuclear charge and current operators consist of one- and two - body terms that operate on the nucleon degrees of freedom : @xmath58 the one - body operators @xmath59 and @xmath60 have the standard expressions obtained from a relativistic reduction of the covariant single - nucleon current , and are listed below for convenience . the charge operator is written as @xmath61 with @xmath62 @xmath63 where @xmath64 is the four - momentum transfer defined earlier , and @xmath65 is the nucleon mass . the current operator is expressed as @xmath66_+ -{\frac { { \rm i } } { 2 m } } \mu_i \
> { \bf q } \times \mbox{\boldmath$ \sigma$}_i \ > { \rm e}^{{\rm i } { \bf q } \cdot { \bf r}_i } \label{eq9}\>\>\>,\ ] ] where @xmath67_+$ ] denotes the anticommutator .
the following definitions have been introduced : @xmath68 \>\>,\label{eq8a}\\ \mu_i & \equiv & \frac { 1}{2}\left [ g_m^s(|q^2| ) + g_m^v(|q^2|)\tau_{z , i } \right ] \label{eq9a } \>\>,\end{aligned}\ ] ] and @xmath69 , @xmath70 , and @xmath71 are the nucleon s momentum , pauli spin and isospin operators , respectively .
the two terms proportional to @xmath72 in @xmath73 are the well known darwin - foldy and spin - orbit relativistic corrections @xcite , respectively .
the dipole parametrization is used for the isoscalar ( @xmath74 ) and isovector ( @xmath75 ) combinations of the electric and magnetic nucleon form factors ( including the galster form for the electric neutron form factor @xcite ) .
the most important features of the two - body parts of the electromagnetic current operator are summarized below .
the reader is referred to refs .
@xcite for a derivation and listing of their explicit expressions .
the two - body current operator has `` model - independent '' and `` model - dependent '' components , in the classification scheme of riska @xcite .
the model - independent terms are obtained from the two - nucleon interaction ( in the present study the argonne @xmath4 interaction @xcite is employed ) , and by construction satisfy current conservation with it .
the leading operator is the isovector `` @xmath76-like '' current obtained from the isospin - dependent spin - spin and tensor interactions .
the latter also generate an isovector `` @xmath77-like '' current , while additional model - independent isoscalar and isovector currents arise from the isospin - independent and isospin - dependent central and momentum - dependent interactions .
these currents are short - ranged and numerically far less important than the @xmath76-like current .
the model - dependent currents are purely transverse and therefore can not be directly linked to the underlying two - nucleon interaction .
the present calculation includes the isoscalar @xmath78 and isovector @xmath79 transition currents as well as the isovector current associated with excitation of intermediate @xmath80-isobar resonances ( for the values of the various coupling constants and cutoff masses in the monopole form factors at the meson - baryon vertices , see ref .
@xcite ) . among the model - dependent currents , those associated with the @xmath80-isobar are the most important ones . in the present calculation , these currents are treated within the static @xmath80 approximation . while this is sufficiently accurate for our purposes here , it is important to realize that such an approach can lead to a gross overestimate of @xmath80 contributions in electro - weak transitions ( see refs .
@xcite for a discussion of this issue within the context of neutron and proton radiative captures on deuteron and @xmath81he , and the proton weak capture on @xmath81he ) .
finally , it is worth pointing out that the contributions associated with the @xmath78 , @xmath79 and @xmath80-excitation mechanisms are , in the regime of low to moderate momentum - transfer values of interest here ( @xmath82 @xmath83 ) , typically much smaller than those due to the leading model - independent @xmath76-like current @xcite .
while the main parts of the two - body currents are linked to the form of the two - nucleon interaction through the continuity equation , the most important two - body charge operators are model - dependent , and should be considered as relativistic corrections .
indeed , a consistent calculation of two - body charge effects in nuclei would require the inclusion of relativistic effects in both the interaction models and nuclear wave functions .
there are nevertheless rather clear indications for the relevance of two - body charge operators from the failure of the impulse approximation in predicting the deuteron tensor polarization observable @xcite , and charge form factors of the three- and four - nucleon systems @xcite .
the model commonly used @xcite includes the @xmath76- , @xmath77- , and @xmath42-meson exchange charge operators with both isoscalar and isovector components , as well as the ( isoscalar ) @xmath78 and ( isovector ) @xmath79 charge transition couplings , in addition to the single - nucleon darwin - foldy and spin - orbit relativistic corrections .
the @xmath76- and @xmath77-meson exchange charge operators are constructed from the isospin - dependent spin - spin and tensor interactions ( those of the argonne @xmath4 here ) , using the same prescription adopted for the corresponding current operators . it should be emphasized , however , that for @xmath84 @xmath83 the contributions due to these two - body charge operators are very small when compared to those from the one - body operator . in the standard model the vector part of the neutral weak current
is related to the isoscalar ( @xmath74 ) and isovector ( @xmath75 ) components of the electromagnetic current , denoted respectively as @xmath85 and @xmath86 , via @xmath87 and therefore the associated one- and two - body weak charge and current operators are easily obtained from those given in the preceding section . the axial charge and current operators too have one- and two - body terms .
only the axial current @xmath88 enters in the calculation of the asymmetry .
the axial charge operator is not needed in the present work .
the one - body axial current is given , to lowest order in @xmath89 , by @xmath90 where the nucleon axial form factor is parametrized as @xmath91 here @xmath92 is the nucleon axial coupling constant , @xmath93 , and the cutoff mass @xmath94 is taken to be 1 gev / c@xmath2 , as obtained from an analysis of pion electroproduction data @xcite and measurements of the reaction @xmath95 @xcite .
there are relativistic corrections to @xmath96 as well as two - body contributions arising from @xmath76- , @xmath77- , @xmath97-exchange mechanisms and @xmath80 excitation @xcite .
all these effects , however , are neglected in the present study .
the reasons for doing so are twofold : firstly , axial current contributions to the asymmetry are small , since they are proportional to the electron neutral weak coupling @xmath98 ( see eq . ( [ eq : asy ] ) ) ; secondly , axial contributions from two - body operators are expected to be at the @xmath99 % level of those due to the one - body operator in eq .
( [ eq : axc1 ] ) .
for example , in the proton weak capture on proton at kev energies @xcite
this process is induced by the charge - changing axial weak current the @xmath76 , @xmath77 , @xmath97 , and
@xmath80 two - body operators increase the predicted one - body cross section by 1.5% .
such an estimate is expected to hold up also in the quasi - elastic regime being considered here .
the purpose of this section is to relate the operators presented in sections [ sec : emop ] and [ sec : weop ] to chpt .
we consider only up and down quarks , in which case the qcd lagrangian has an approximate @xmath100 chiral symmetry .
this symmetry is spontaneously broken by the vacuum to the diagonal @xmath101 subgroup , and three pseudoscalar goldstone bosons , the pions @xmath102 , appear in the spectrum .
chiral symmetry provides important constraints on the description of low - momentum processes involving pions .
in particular , it allows one to estimate the relative size of various contributions . to accomplish this ,
the most general effective lagrangian with broken @xmath100 is constructed .
this effective lagrangian includes terms with arbitrary number of derivatives and powers of the quark masses , however higher - dimension operators are suppressed by inverse powers of the characteristic mass scale of qcd , @xmath103 gev .
thus pion interactions are determined as a power series in @xmath104 , where @xmath41 is the typical external three - momentum . at low energies
this is a small number and hence only the lowest - order terms are considered in this paper .
we start with the @xmath105 and @xmath106 lagrangians .
details can be found , for example , in refs .
we lump the pions in a field @xmath107 , where @xmath108 mev is the pion decay constant .
we denote the nucleon field of velocity @xmath109 and spin @xmath110 by @xmath7 . in order to obtain the currents required for the computation of the asymmetry we must use the proper covariant derivatives .
the covariant derivatives on the pion and nucleon field are constructed in terms of external vector and axial - vector fields @xmath111 and @xmath112 in the usual manner .
they are given by @xmath113 @xmath114n - \frac{\rm i}{2}\bigg [ u^{\dagger}({\cal v}_{\mu}^{(3)}+{\cal a}_{\mu}^{(3)})u + u ( { \cal v}_{\mu}^{(3)}-{\cal a}_{\mu}^{(3)})u^{\dagger}\bigg ] n -3{\rm i}\,{\cal v}_{\mu}^{(0)}n,\ ] ] where the superscripts @xmath115 and @xmath116 denote isoscalar and isovector components .
it is convenient to construct also other quantities that transform covariantly .
for example , @xmath117_+ , \\ f_{\mu \nu}^\dagger & = & u^\dagger f_{\mu \nu}^ru + uf_{\mu \nu}^lu^\dagger , \end{aligned}\ ] ] where @xmath118 $ ] with @xmath119 and @xmath120 .
one can find the relation between the external fields and @xmath0 or photon by considering the covariant derivative on the quark fields ( see , e.g. , ref .
@xcite ) : @xmath121 @xmath122 where @xmath123 is the quark - charge matrix . from these building blocks we can write the chiral lagrangians , @xmath124 + \ldots , \\\ ] ] @xmath125n+ \frac{1}{2 m } n^\dagger\bigg[(v\cdot d)^2-d\cdot d - { \rm i}\,g_a [ s\cdot d , v\cdot v]_{+ } \bigg ]
n \nonumber\\ -\frac{\rm i}{4m}n^\dagger [ s^\mu , s^\nu ] \bigg[(1-\kappa_v ) f_{\mu \nu}^\dagger + \frac{1}{2}(\kappa_s-\kappa_v){\rm tr}(f_{\mu \nu}^\dagger)\bigg]n + \ldots,\end{aligned}\ ] ] where @xmath126 and @xmath127 , and @xmath128 denote terms with more derivatives and/or powers of the pion mass .
one can also write down interactions containing four or more nucleon fields @xcite , which are important for a fully consistent description of systems involving two or more nucleons .
one and two - body currents can be obtained from these interactions .
the symmetries allow an infinite number of interactions , so an ordering scheme is necessary for predictive power .
we want to estimate the size of matrix elements of one- or two - body currents between @xmath7@xmath7 wave functions .
these matrix elements involve : the final @xmath7@xmath7 wave function , a two - nucleon propagator , the current operator , another two - nucleon propagator , and the initial @xmath7@xmath7 wave function .
each two - nucleon propagator brings a factor @xmath129 , and each loop a @xmath130 .
apart from loops in the operators themselves , matrix elements of one- and two - body operators involve , respectively , one and two loops .
let us first consider only strong interactions between nucleons @xcite .
these start at @xmath131 , and most structures that appear up to @xmath132 @xcite are present in the argonne @xmath4 potential @xcite .
use of phenomenological potentials in conjunction with currents derived in chpt has already proven to be a successful approach @xcite .
contributions to the amplitude @xmath133 start at @xmath134 with the tree - level one - body charge operator .
first corrections come in tree - level one - body currents from @xmath135 magnetic corrections in the lagrangian .
second corrections are of two types : ( i ) one - loop corrections and @xmath136 interactions in one - body currents ; and ( ii ) tree - level two - body currents .
and so on .
contributions to the amplitude @xmath29 follow the same pattern , but have an extra overall factor of @xmath137 .
the leading one and two - body contributions are described in sec .
[ sec : chptob ] below .
electromagnetic interactions between nucleons are clearly higher - order effects , also included in the @xmath4 potential .
interesting are also the contributions to the asymmetry that come not from the exchange of a @xmath0 between electron and deuteron , but from @xmath0 exchange between hadrons ( and photon exchange between electron and deuteron ) .
let us denote these contributions by @xmath138 .
how does @xmath138 compare to @xmath29 ?
direct @xmath139 exchange between nucleons is an @xmath140 effect , so suppressed by @xmath141 compared to the strong interaction . on the other hand ,
the electron - deuteron interaction is @xmath142 larger in @xmath138 than in @xmath29 . as a consequence , the contribution from @xmath0 exchange between nucleons to @xmath138
is @xmath143 compared to the leading contribution to @xmath29 .
since we are interested in @xmath144 , this is @xmath145 , a small effect .
yet , we are here searching for subleading contributions .
as we have seen , two - body currents in @xmath29 are suppressed by @xmath136 compared to the leading one - body effect .
thus , compared to two - body currents in @xmath29 , @xmath138 can be @xmath146 ; for momenta @xmath144 , this is @xmath147 .
there exist also potentially larger contributions to @xmath138 .
the size of the parity - violating pion - nucleon coupling might be as large as @xmath148 compared to the parity - conserving coupling @xcite .
in this case pion exchange in the one - body current ( anapole form factor ) , in the potential and in the two - body current could all produce an effect @xmath149 compared to two - nucleon effects in direct @xmath0 exchange .
our calculation involves all contributions suppressed by @xmath136 ( and also some even more suppressed ) compared to leading , except for @xmath138 .
however , contributions to @xmath138 from the parity - violating potential were already examined in ref . @xcite . also ,
although not included in any deuteron calculation , the anapole form factor of the nucleon seems too small @xcite to be relevant here , and it is reasonable to expect that two - body currents stemming from the parity - violating pion - nucleon coupling will be equally unimportant .
( this should be eventually confirmed in an explicit calculation . ) on the other hand , the leading two - body currents in the amplitude @xmath29 stemming from pion exchange are being calculated here for the first time
. .25 cm .75 cm the one - body currents contributing to the processes in fig .
[ fig : ampl ] , depicted in figs .
[ f1 ] and [ f2 ] , are given to @xmath150 by @xmath151 \nonumber \\ & & + ( g_m^s+g_m^v\tau_{z , i } ) \mbox{\boldmath$ \nabla$}\delta({\bf r}-{\bf r}_i)\times \mbox{\boldmath$ \sigma$}_i \bigg]\\ \rho^5({\bf r } ) & = & -\frac{1}{4m}g_a\sum_{i } [ \mbox{\boldmath$ \sigma$}_i\cdot { \bf p}_i \delta({\bf r}-{\bf r}_i)+ \delta({\bf r}-{\bf r}_i)\mbox{\boldmath$ \sigma$}_i\cdot { \bf p}_i ] \tau_{z , i } \\ { \bf j}^5({\bf r } ) & = & -\frac{1}{2 } g_a\sum_{i } \delta({\bf r}-{\bf r}_i)\mbox{\boldmath$ \sigma$}_i \tau_{z , i}.\end{aligned}\ ] ] where @xmath152 or @xmath54 , the coupling constants @xmath153 and @xmath154 are given in table [ table1 ] , and @xmath92 is the nucleon axial coupling constant .
these results are in agreement with those presented in ref .
these coupling constants acquire , in higher orders , a @xmath30 dependence , see ref . @xcite . in our calculation
we use a phenomenological parametrization of the @xmath30 dependence , as described in secs .
[ sec : emop ] and [ sec : weop ] . .coupling constants appearing in one - body currents to @xmath150 in chpt . [ cols="^,^,^",options="header " , ] note that @xmath155 contains a @xmath76-@xmath0 mixing term in the form @xmath156 , which turns out to be proportional to @xmath157 , the four - momentum transfer .
its contraction with the leptonic current produces a contribution proportional to the mass of the electron . in the extreme relativistic limit for the electron under consideration here , this contribution can be neglected . .25 cm .25 cm the two - body contributions to the processes shown in fig .
[ fig : ampl ] are depicted in figs .
[ f3 ] and [ f4 ] , where again contributions from @xmath76-@xmath0 mixing are neglected in the extreme relativistic limit . to @xmath158 , they are given in momentum space by @xmath159 \ , \end{aligned}\ ] ] where @xmath160 with @xmath161 ( @xmath162 ) denoting the initial ( final ) momentum of nucleon @xmath163 . in the formula above , @xmath164 note that there is no contribution to the axial current operator or the electromagnetic charge operator .
there is a contribution to the axial charge operator which , however , does not enter in the asymmetry computation .
these results are in agreement with ref .
@xcite , where the fourier transform of the above expressions are also given in detail . in higher order
other currents appear .
there exist shorter - range currents , which are expected to be smaller than the ones from pion exchange with leading order interactions .
these higher - order effects are parametrized in our calculation through the riska prescription as outlined in sections [ sec : emop ] and [ sec : weop ] .
in particular , we do not use eq .
( [ vpichpt ] ) for @xmath165 , but the pseudoscalar component of the @xmath4 potential .
in this section we describe the calculation of the deuteron response functions given in eqs .
( [ eq : rl])([eq : rtp ] ) .
the deuteron wave function is written as @xmath166\chi_0 ^ 0 \ , \label{eq : deu}\ ] ] where the @xmath167 are standard spin - angle functions , @xmath168 is a two - nucleon @xmath169 isospin state , and @xmath170 and @xmath171 are the s- and d - wave radial functions . in the @xmath2h(@xmath172)@xmath3 reaction the final state is in the continuum , and
its wave function is written as @xmath173 where @xmath174 and @xmath175 are the relative and center - of - mass coordinates .
the incoming - wave scattering - state wave function of the two nucleons having relative momentum @xmath69 and spin - isospin states @xmath176 is approximated as @xcite @xmath177 \chi_{m_s}^s \chi_{m_t}^t + { 4 \pi \over \sqrt{\ , 2 } } \sum_{\scriptstyle jm_j\atop\scriptstyle j\leq j_{\rm max } } \sum_{ll ' } { \rm i}^l\ , \delta_{lst } \nonumber \\ & & \quad \bigl [ z_{lsm_s}^{jm_j}(\hat { \bf p})\bigr]^ * \biggl [ { 1 \over r } u_{l'l}^{(- ) } ( r ; p , jst ) - \delta_{l'l}\>\ > j_l ( pr ) \biggr]\ , { \cal y}_{l'sj}^{m_j } \chi_{m_t}^t \ , \label{eq : pnsc}\end{aligned}\ ] ] where @xmath178 the @xmath179 factor ensures the antisymmetry of the wave function , while the clebsch - gordan coefficients restrict the sum over @xmath180 and @xmath181 .
the radial functions @xmath182 are obtained by solving the schrdinger equation in the @xmath183 channel , and behave asymptotically as @xmath184 \>\>,\ ] ] where @xmath185 is the @xmath74-matrix in the @xmath183 channel and the hankel functions are defined as @xmath186=@xmath187 , @xmath188 and @xmath189 being the spherical bessel and neumann functions , respectively . in the absence of interactions , @xmath190 , and @xmath191 reduces to an antisymmetric plane wave .
interactions effects are retained in all partial waves with @xmath192 . in the quasi - elastic regime of interest here , it is found that these interaction effects are negligible for @xmath193 .
the response functions are written as ( only @xmath194 is given below for illustration ) @xmath195 where the contributions from the individual spin - isospin states are @xmath196 with @xmath197 defined as @xmath198 here @xmath199 mev is the deuteron ground - state energy , the factor 1/2 in eq .
( [ eqrst ] ) is included to avoid double counting , and the states @xmath200 and @xmath201 are represented by the wave functions in eqs .
( [ eq : deu ] ) and ( [ eq : pnsc ] ) , respectively . by integrating out the energy - conserving @xmath202-function
one finds : @xmath203 where the magnitude of the relative momentum @xmath69 is fixed by @xmath204 , and @xmath205 is the angle between @xmath56 and @xmath69 .
the initial- and final - state wave functions are written as vectors in the spin - isospin space of the two nucleons for any given spatial configuration @xmath206 .
for the given @xmath206 the state vector @xmath207 is calculated with the same methods used in quantum monte carlo calculations of , for example , the charge and magnetic form factors of the trinucleons @xcite .
the @xmath206 and @xmath205 integrations required to calculate the amplitudes and response function are then performed by means of gaussian quadratures .
finally , note that , since the deuteron is a @xmath208 state , one finds @xmath209 with similar relations holding between the transverse response functions .
the asymmetry has been calculated at the kinematics relevant to the sample experiment .
the incident electron energy was set to @xmath210 mev .
sample measures the asymmetry at four different angles ( @xmath211 ) .
different electron final energies @xmath212 correspond to different momentum transfers @xmath30 , @xmath1 gev@xmath2 in the sample experiment , which is small enough to justify the use of a non - relativistic formalism with leading interactions obtained from chpt .
the calculated asymmetries , as functions of the electron final energy , are shown in figs .
[ f10 ] , [ f11 ] , [ f12 ] , and [ f13 ] , for the four different electron scattering angles .
for each set of kinematics , the left panels display the asymmetry and the total inclusive cross section , with different curves representing one - body contributions , one- plus two - body contributions from pion - exchange currents only , and the sum of all contributions .
the ratios of one- plus two - body contributions from pion only and full currents to one - body contributions for both asymmetries and cross sections are displayed in the right panels . .25 cm .25 cm .25 cm .25 cm as is apparent from the figures , the results at all angles are qualitatively similar . near the quasi - elastic peak two - body effects in the asymmetry
are negligible , less than 1% , while away from the quasi - elastic peak they become relatively more important , increasing the asymmetry by at the most 3% .
note , however , that the two - body current contributions are large in the inclusive cross section , indeed dominant in the left - hand side of the quasielastic peak . in this region
the contribution associated with the currents of pion range is more than 50% of the total two - body contribution .
it is interesting to examine more closely the reasons for the relative unimportance of two - body current contributions in the asymmetry . at backward angles ,
the expression for the asymmetry can be approximated as @xmath213 where terms proportional to the longitudinal response functions are suppressed by the factor @xmath214 , a small number at the angles under consideration here ( note that the @xmath215 and @xmath216 response functions are of the same order of magnitude ) .
it is useful to identify the contributions from @xmath208 and @xmath217 @xmath3 final states , and to use eqs .
( [ eq : iso1 ] ) and ( [ eq : iso2 ] ) , relating the @xmath218 to @xmath219 .
one then finds : @xmath220 note that the @xmath221 response function only receives contributions from @xmath217 @xmath3 final states , since the current @xmath96 is isovector .
the ratio @xmath222 is much smaller than one , since the transverse response is predominantly isovector .
for example , at @xmath223 mev and @xmath224 , @xmath225 mev@xmath226 and @xmath227 mev@xmath226 , and hence @xmath228 with one - body ( full ) currents .
in contrast , the ratio @xmath229 is of order one ; again at @xmath223 mev and @xmath224 , @xmath230 mev@xmath226 , and hence @xmath231 with one - body ( full ) currents .
however , it is multiplied by the small factor @xmath232 , and so the asymmetry turns out to be largely independent of nuclear structure details . finally , if @xmath233 denotes the asymmetry obtained by ignoring the contribution of the axial current , one finds @xmath234 the computed value for this ratio is shown in fig . [ f14 ] for one of the kinematics of the sample experiment .
the contribution of the axial current to the asymmetry is of the order of 13% to 24% throughout the kinematical range considered
. note that in fig .
[ f14 ] we have included the small contribution from the longitudinal response . as a last remark , we should emphasize that the calculated transverse ( @xmath235 ) and longitudinal ( @xmath194 ) response functions including one- and two - body operators reproduce @xcite existing bates data @xcite .
.25 cm since we have performed the computations at the sample kinematics , the above results may be used to account for two - body current corrections in the analysis of the experimental data .
the sample experiment measures a convolution of the asymmetry @xmath236 and the cross section @xmath237 over a certain range of electron final energies : @xmath238 the goal of the experiment is to extract the one - body part of @xmath239 . to accomplish this , a model that includes only one - body contributions
is used to generate the cross section .
one can now use our results for the ratios of total to one - body contributions in the asymmetry and cross section to adjust for two - body effects in the experiment . in conclusion
, we have presented a fairly complete calculation of the asymmetry in quasi - elastic electron - deuteron scattering arising from @xmath0 exchange .
since we find that , when the cross section is large at the quasi - elastic peak , the change in the asymmetry due to two - body currents is negligible , we expect that these two - body corrections will produce a modification in the analysis of the sample experiment at the % level , too small to affect significantly the extraction of the strange and axial form factors of the nucleon .
it remains to be examined whether the same holds for effects from @xmath0 exchange that manifest themselves within the two - nucleon system through the parity - violating pion - nucleon coupling .
work along these lines is in progress .
we are grateful to bob mckeown for encouragement and many useful suggestions .
l.d . thanks the jefferson laboratory for hospitality .
r.s . thanks the kellogg radiation laboratory staff and , in particular , bob mckeown for the warm hospitality extended to him during his january visits in the past three years .
u.vk . thanks betsy beise and takeyasu ito for discussions regarding the sample experiments . the work of l.d . and u.vk . was supported in part by nsf grant phy 94 - 20470 .
the work of r.s . was supported by doe contract de - ac05 - 84er40150 under which the southeastern universities research association ( sura ) operates the thomas jefferson national accelerator facility .
u.vk . thanks riken , brookhaven national laboratory and the u.s .
department of energy under contract de - ac02 - 98ch10886 for providing the facilities essential for the completion of this work .
finally , some of the calculations were made possible by grants of computing time from the national energy research supercomputer center in livermore .
( sample collaboration ) , phys .
84 * , 1106 ( 2000 ) .
e. beise and m. pitt ( spokespersons ) , mit - bates proposal .
mckeown , private communication ( 2000 ) .
s. ying , w.c .
haxton , and e.m .
henley , phys .
d * 40 * , 3211 ( 1989 ) .
e. hadjimichel , g.i .
poulis , and t.w .
donnelly , phys .
c * 45 * , 2666 ( 1992 ) .
s. schram and c.j .
horowitz , phys .
c * 49 * , 2777 ( 1994 ) .
hwang , e.m .
henley , and g.a .
miller , ann .
( n.y . ) * 137 * , 378 ( 1981 ) .
v. bernard , n. kaiser , and u .- g .
meiner , int .
e4 * , 193 ( 1995 ) .
u. van kolck , prog .
. phys . * 43 * , 409 ( 1999 ) ; s.r .
et al . _ , nucl - th/0008064 .
j. carlson and r. schiavilla , rev .
phys . * 70 * , 1 ( 1998 ) .
riska , phys . rep . * 181 * , 207 ( 1989 ) .
wiringa , v.g.j .
stoks , and r. schiavilla , phys .
c * 51 * , 38 ( 1995 ) .
t.m . ito ( spokesperson ) , mit - bates proposal .
g0 collaboration , www.npl.uiuc.edu/exp/g0/g0main.html m.j .
musolf and t.w .
donnelly , nucl .
* a546 * , 509 ( 1992 ) . j.l .
friar , ann .
( n.y . ) * 81 * , 332 ( 1973 ) .
s. galster _ et al .
_ , nucl .
* b32 * , 221 ( 1971 ) .
r. schiavilla , v.r .
pandharipande , and d.o .
riska , phys .
c * 40 * , 2294 ( 1989 ) .
r. schiavilla , v.r .
pandharipande , and d.o .
riska , phys .
c * 41 * , 309 ( 1990 ) .
j. carlson , d.o .
riska , r. schiavilla , and r.b .
wiringa , phys .
c * 42 * , 830 ( 1990 ) .
m. viviani , a. kievsky , l.e .
marcucci , s. rosati , and r. schiavilla , phys . rev .
c * 61 * , 064001 ( 2000 ) .
m. viviani , r. schiavilla , and a. kievsky , phys .
c * 54 * , 534 ( 1996 ) .
r. schiavilla , r.b .
wiringa , v.r .
pandharipande , and j. carlson , phys .
c * 45 * , 2628 ( 1992 ) .
marcucci _ et al .
_ , nucl - th/0006005 .
marcucci , d.o .
riska , and r. schiavilla , phys .
c * 58 * , 3069 ( 1998 ) .
d. abbott _ et al .
_ , phys .
84 * , 5053 ( 2000 ) .
wiringa , phys .
c * 43 * , 1585 ( 1991 ) .
e. amaldi , s. fubini , and g. furlan , _ electroproduction at low energy and hadron form factors _ ( springer tracts in modern physics no . 83 ) .
t. kitagaki _ et al .
_ , phys .
d * 28 * , 436 ( 1983 ) .
r. schiavilla _ et al .
_ , phys .
c * 58 * , 1263 ( 1998 ) .
m.e . peskin and d.v schroeder , _ an introduction to quantum field theory _
( addison - wesley , 1997 ) . c. ordez , and u. van kolck , phys
b291 * , 459 ( 1992 ) ; l. ray , c. ordez , and u. van kolck , phys .
c * 53 * , 2086 ( 1996 ) .
maekawa and u. van kolck , phys .
* b478 * , 73 ( 2000 ) ; c.m .
maekawa , j.s .
veiga , and u. van kolck , phys . lett . *
b488 * , 167 ( 2000 ) .
v. bernard , h.w .
fearing , t.r .
hemmert , and u .- g .
meiner , nucl . phys . *
a635 * , 121 ( 1998 ) ; * a642 * , 563 ( 1998 ) ( e ) .
r. schiavilla and d.o .
riska , phys .
c * 43 * , 437 ( 1991 ) .
j. carlson and r. schiavilla , phys .
* 68 * , 3682 ( 1992 ) .
s.a . dytman _
et al . _ ,
c * 38 * , 800 ( 1988 ) . | the longitudinal asymmetry due to @xmath0 exchange is calculated in quasi - elastic electron - deuteron scattering at momentum transfers @xmath1 gev@xmath2 relevant for the sample experiment .
the deuteron and @xmath3 scattering - state wave functions are obtained from solutions of a schrdinger equation with the argonne @xmath4 potential .
electromagnetic and weak neutral one- and two - nucleon currents are included in the calculation .
the two - nucleon currents of pion range are shown to be identical to those derived in chiral perturbation theory .
the results indicate that two - body contributions to the asymmetry are small ( @xmath5 0.2% ) around the quasi - elastic peak , but become relatively more significant ( @xmath5 3% ) in the high - energy wing of the quasi - elastic peak .
= 0.25 in = 0.25 in 16.cm = .10 in = 1ex = 2ex = 6ex = 3ex = -0.3 in plus 2pt minus 2pt * l. diaconescu * + and + _ department of physics , old dominion university , norfolk , virginia 23529 _ + + and + _ riken - bnl research center , brookhaven national laboratory , upton , ny 11973 _ + and + _ kellogg radiation laboratory , california institute of technology , pasadena , ca 91125 _ pacs : 13.10.+q , 21.45.+v , 25.30.fj |
one of the strange features of high - temperature cuprate superconductors , a normal - state gap in the density of states , remains a subject of controversy .
a great deal of experimental evidence ( nmr,@xcite optical conductivity,@xcite specific heat,@xcite and , most recently , angle - resolved photoemission@xcite ) seems to indicate that the superconducting energy gap survives the transition to the normal state and disappears only at a considerably higher temperature .
even the anisotropic character of the gap is preserved .
this behavior is characteristic of `` underdoped '' compounds , in which doping of additional hole carriers increases the temperature of the superconducting transition @xmath1 .
there is no generally accepted theoretical model of this phenomenon , although arguably the most popular tentative explanation puts the blame on superconducting fluctuations near the transition point , which are expected to be enhanced in these highly anisotropic , almost two - dimensional , materials .
while monte - carlo simulations of the attractive hubbard model in two dimensions provide numerical evidence@xcite for the pairs - above-@xmath1 scenario , no analytical description of the pseudogap regime has been offered so far .
the goal of this paper is to provide such a sketch .
suppose that the pseudogap in cuprates does have the same origin as the superconducting gap the scattering between fermion and hole states with the charge @xmath2 released in the form of a low - energy cooper pair .
this picture rests on certain assumptions . simply saying that cooper pairs are present above @xmath1 as a propagating mode is not enough .
it is important that the fluctuating pairing field @xmath3 look frozen on the time scale @xmath4 , otherwise the pseudogap will not be formed . since the typical frequency of free propagating bosons in thermal equilibrium
is given by their temperature , the pseudogap is expected to exist at temperatures @xmath5 , where @xmath6 is loosely defined to be of order @xmath7 . in general
, it may also be necessary to require the _ spatial _ coherence of the pairing field over a pair size @xmath8 . in this particular model , however , the boson velocity is of order of the fermi velocity @xmath9 , so that as long as the pairing fluctuations are slow in time , they are also slow in space . in a somewhat arbitrary way , @xmath6 can be identified with the transition temperature calculated in the bcs theory . as in any other mean - field theory ,
the onset of a long - range order coincides with the appearance of a short - range order ( formation of pairs ) , the assumption being that the phase of the order parameter will lock up as soon as there is a non - zero amplitude .
this , of course , breaks down in two spatial dimensions , as shown long ago by hohenberg.@xcite in a highly anisotropic , almost two - dimensional system , long - range order sets in at a temperature @xmath10 . above @xmath6 , cooper pairs decay
; between @xmath6 and @xmath1 , they represent a propagating mode ; finally , below @xmath1 , they form a bose condensate . in the intermediate range ,
at least when @xmath11 , they represent a slowly fluctuating bcs pairing field and thus create a pseudogap . a well - defined pseudogap regime may be a peculiarity of low - dimensional systems .
a theoretical framework for treating the interplay between fermions and their boson - like bound states is thus needed . following pioneering works by eagles,@xcite leggett,@xcite and nozieres and schmitt - rink,@xcite several approximate methods
have been suggested : the boson - fermion model,@xcite functional integration,@xcite and the self - consistent @xmath0-matrix approximation.@xcite in most cases , however , one has to resort to numerical computations .
the present work is written with the purpose to provide an analytical sketch of a _ normal state with a pseudogap_. as such , it inevitably contains further simplifications , which hopefully do not alter the nature of the problem : ( 1 ) the decay of low - energy bosons is neglected .
( 2 ) a clear separation of energy scales is assumed : @xmath12 ( @xmath13 is the fermi energy ) . using the self - consistent @xmath0-matrix approach ,
i derive approximate propagators for fermions and cooper pairs .
the width of the pseudogap is determined by the mean - square fluctuation of the pairing field @xmath14 in a thermal ensemble .
this expectation value depends , among other things , on the energy spectrum of cooper pairs , which in turn is a function of the fermion energy spectrum . a closed set of coupled equations results , allowing one to determine the condensation temperature of pairs .
the so determined @xmath1 is proportional to the fermion density and inverse mass , as seen on the uemura@xcite plot .
this is despite the fact that , as long as @xmath15 , the system is not in the limit of local ( tightly bound ) pairs .
i also give a reason why quasiparticle peaks are quite broad near the fermi surface , even when thermal fluctuations are slow ( @xmath16 ) . if correct , this sketch may provide a simple way to understand two puzzling features of underdoped cuprate superconductors the pseudogap and the doping dependence of @xmath1 from a unified standpoint .
the paper is organized as follows .
the conserving @xmath0-matrix approximation is outlined in section [ t - matrix ] ; a bosonic propagator for a cooper pair is introduced for the case of a `` separable '' interaction vertex .
section [ fermions ] contains a derivation of the fermion propagator in the presence of a slowly fluctuating pairing field .
an exact model of section [ gaussian fluctuations ] reveals an important difference between a true superconducting gap and a pseudogap , stemming from the quantum nature of pairing fluctuations .
low - energy properties of cooper pairs in the presence of a pseudogap and their condensation temperature are derived in section [ bosons ] .
the effect of fast fluctuations is estimated in the appendix .
a good description of the self - consistent @xmath0-matrix approximation can be found in ref .
[ haussmann ] .
the relation to other approximate methods is outlined in ref .
[ micnas ] .
the conserving @xmath0-matrix approximation is somewhat similar to the well - known hartree - fock principle.@xcite the latter neglects any correlations between interacting particles ( except for statistical ones ) and describes the motion of independent entities in a self - consistent potential .
the @xmath0-matrix approach includes pairwise correlations between colliding particles , which are particularly important when two - particle bound states are formed .
the form of these correlations is inferred from the exact solution of a similar problem with two particles in vacuum , when the two - particle green s function @xmath17\rangle\ ] ] can be expressed in terms of the one - particle green s function @xmath18\rangle$ ] and the two - body @xmath0 matrix [ fig .
[ graphs](a ) ] . in the many - body case ,
the latter is _ defined _ as a solution of the bethe - salpeter equation [ fig .
[ graphs](b ) ] : @xmath19 here @xmath20 is the total 4-momentum of two fermions , @xmath21 and @xmath22 are relative 4-momenta , and @xmath23 is the four - dimensional volume ( with an imaginary time dimension ) . by reducing @xmath24 to a functional of @xmath25 ,
one breaks the infinite hierarchy of equations for @xmath26-particle green s functions because the resulting approximate dyson equation @xmath27 contains no higher - order green s functions [ fig .
[ graphs](c ) ] : @xmath28 equations
[ t ] and [ sigma ] form a closed set of equations that could be solved , at least in principle .
@xcite the use of full fermion propagators ( instead of bare ones ) in these equations is the main difference from the earlier approach of thouless,@xcite equivalent to the bcs theory .
in addition to having conservation laws built in,@xcite the self - consistent approach is a step towards including the influence of gaussian fluctuations on the phase transition . near the transition temperature
, fermions start to feel the presence of the emerging bosonic excitations , which in turn influences the process of pair formation .
dressing of the fermionic propagators in eq .
[ t ] allows one to account for this feedback , at least to some extent .
the use of the dressed fermion propagator in the self - energy equation ( [ sigma ] ) increases the number of terms in the perturbation series for @xmath29 , which turns out to be important for treating properly the quantum nature of fluctuating cooper pairs .
although the @xmath0 matrix depends on three independent momenta , the non - trivial dependence should be associated with the conserved total momentum @xmath30 .
it is convenient to write the interaction vertex formally as [ fig .
[ graphs](d ) ] @xmath31 where @xmath32 and @xmath33 are some relative coordinates .
one well - known example of such factorization is the case of a separable instanteneous potential @xmath34 .
this formalism is particularly convenient for lattice models with finite - range instanteneous interactions . in the case of the @xmath35-@xmath36 model , for instance ,
@xmath37 in the singlet channel and zero in the triplet one .
the infinite on - site repulsion imposes the constraint of no double occupancy .
phonon - mediated attraction can also be written in the form ( [ factorization ] ) with @xmath38 .
the @xmath0 matrix can now be found using the _ ansatz _ @xmath39 [ fig .
[ graphs](e ) ] . the new matrix @xmath40 satisfies the dyson equation @xmath41 with the polarization matrix @xmath42 clearly
, @xmath40 , which contains all the non - trivial information about the @xmath0 matrix , can be regarded as the bosonic propagator of a cooper pair whose internal structure is described by the dependence on the relative coordinates @xmath32 and @xmath33 .
the factorization procedure described above allows one to write the bethe - salpeter equation for the @xmath0 matrix ( [ t ] ) in the form reminiscent of the dyson equation for a doubly charged bosonic particle see fig .
[ graphs](f ) without introducing spurious degrees of freedom .
the thouless criterion for pair condensation , @xmath43 at @xmath44 , simply means that the quasiparticles with @xmath45 have zero excitation energy .
the fermion self - energy ( [ sigma ] ) can now be rewritten using the boson propagator @xmath46 [ fig .
[ graphs](g ) ] : @xmath47 this diagram illustrates two important points about the nature of the @xmath0-matrix approximation . ( 1 ) the irreversible decay of a fermionic quasiparticle is determined by the density of states `` hole + cooper pair '' at a given 4-momentum . note that both outgoing lines are dressed and thus represent actual excitations of the system .
this should be contrasted to the functional - integration approach at the usual one - loop level@xcite where the fermion self - energy is expressed in terms of the _ bare _ fermion propagator .
we will return to this point later in sec .
[ gaussian fluctuations ] .
( 2 ) for a slowly fluctuating pairing field , the sum in eq .
[ sigma bosonized ] is dominated by the region near @xmath44 , so that a fermion with 4-momentum @xmath21 is coupled mostly to the hole with the same 4-momentum . by replacing @xmath48 with @xmath49 , we obtain @xmath50^{-1}},\ ] ] where @xmath51 the approximation made here may appear rather crude .
essentially , the scattering of a fermion into a _ continuum _ of hole states is replaced by the coupling to a _ single _
hole state with the same 4-momentum , a situation reminiscent of the bcs superconductor .
if a bare propagator were used for the hole , the resulting bogoliubov quasiparticles would be stable .
however , the use of a full propagator @xmath49 already makes the lifetime finite . in this case , neglecting temporal and spatial fluctuations of the pairing field does not look so bad , especially in a low - dimensional system .
the approximate equation for the self - energy ( [ slow sigma ] ) can be readily solved.@xcite iterating it once results in a quadratic equation for @xmath52 . since @xmath53 ,
the dressed propagator is @xmath54 the branch of the square root is fixed by the requirement that @xmath55 as @xmath56 . in the simplest case of a short - range instanteneous attractive potential ,
@xmath57 can be replaced by a constant and we have @xmath58 where @xmath59 is the bare fermion energy .
it is readily seen that the spectral weight of the dressed fermion is distributed over two finite regions @xmath60 as follows : @xmath61 this distribution is reminiscent of the smeared bcs peaks at @xmath62 .
in fact , the ratio of the spectral weights ( [ weight in t ] ) at @xmath63 is the same as in the bcs case .
since the spectral weight is pushed away from the fermi surface , a pseudogap opens up in the total density of states @xmath64 ( fig .
[ pseudogap ] ) : @xmath65 where @xmath66 is the density of states in the free case and the function @xmath67 can be expressed in terms of complete elliptic integrals : @xmath68 \mbox { for } x>1 .
\end{array } \right.\ ] ] it vanishes linearly for small values of @xmath32 and approaches 1 as @xmath69 .
the @xmath0-matrix approximation amounts to summing up an infinite number of terms of the perturbation series . yet
an even `` greater '' infinity of terms is left out .
in such a case , when approximations are not easily controlled , it is desirable to check the result using some exactly solvable model .
consider a free fermion gas interacting with a classical external pairing field , which is represented by the action term @xmath70d^4x\ ; d^4x'.\ ] ] the field @xmath71 can be interpreted as the wavefunction of a cooper pair with its constituent fermions at @xmath32 and @xmath33 ; the interaction term above removes two free fermions and forms a bound state .
the @xmath72 symmetry , related to the conservation of charge , makes the phase of the complex field @xmath3 unobservable .
when pair formation lowers the energy of the system , the density of pairs @xmath73 may become large .
there are two distinct possibilities in this case .
the first one occurs when the @xmath72 symmetry is sponateously broken and @xmath3 chooses some direction in the complex plane with the well - known results ( the bcs superconductor with a gap and a bogoliubov sound mode ) .
the second possibility is a symmetric phase with a fluctuating field @xmath3 .
the quantum nature of these fluctuations leads to a non - trivial effect : a significant broadening of the fermion spectral weight .
quantum fluctuations in this model can be implemented by using an ensemble of pairing fields @xmath3 with a symmetric distribution . although every _ single _ measurement of @xmath3 produces a definite non - zero result , @xmath74 upon averaging over a long series of such measurements .
this procedure is feynman s way of quantizing a classical system.@xcite a closely related exactly solvable model for the peierls gap in quasi - one - dimensional metals has been discussed by mckenzie@xcite following the analytic solution of sadovskii.@xcite a generic diagram of the perturbation series for the fermion propagator contains @xmath26 incoming and @xmath26 outgoing dashed lines representing the pairing field [ fig . [ gauss](a ) ] . if @xmath75 , there is also the anomalous fermion propagator , which gives rise to the meissner effect and superconductivity .
consider the extreme case when there are no fluctuations in the ensemble : @xmath76 .
the dressed fermion propagator is simply a geometrical series with a single term in each order . for an instanteneous short - range pairing field @xmath77 ( uniform for simplicity ) , @xmath78 although the geometrical series @xmath79 diverges when @xmath80 , its analytic continuation , the function @xmath81 , has only one singular point @xmath82 .
the bcs result is recovered : @xmath83 the divergence of the perturbation series ( [ broken ] ) in the range @xmath84 is associated with the transfer of the pole of the propagator from @xmath85 to the new locations at @xmath86 .
the structure of the perturbation theory changes considerably in the symmetric phase . assuming that the statistical distribution of the pairing fields is gaussian ,
we can write any average that involves a product of @xmath26 fields @xmath87 and @xmath26 fields @xmath88 as a sum over all possible products of pairwise averages : @xmath89 where @xmath30 represents a permutation of the numbers @xmath90 . upon introducing a dashed line for each pair average @xmath91 , we find @xmath92 different diagrams with @xmath26 such lines [ fig .
[ gauss](b ) and ( c ) ] .
again , take the example of a uniform instantaneous short - range pairing field : @xmath93 each _ single _ field breaks the phase symmetry and produces the effect described above . averaging over the ensemble , however
, restores this symmetry ; the perturbation series now contains @xmath92 identical terms in the @xmath26-th order : @xmath94 clearly , the result is different from ( [ broken ] ) .
expression ( [ restored ] ) was obtained in ref .
[ ren ] in connection with the pseudogap in the boson - fermion model as a leading term in two dimensions .
although the series @xmath95 diverges for any @xmath96 , it represents an asymptotic taylor expansion of the function @xmath97 the divergence of the series is related to the nonanalyticity of @xmath98 on the positive side of the real axis , where it has a cut .
rather then working with a series that converges only in the trivial case @xmath99 , we can sum up the diagrams for a single uniform field @xmath3 as in sec .
[ broken phase ] and then average the result over a gaussian ensemble with the weight @xmath100 : @xmath101 which indeed has ( [ restored ] ) as an asymptotic expansion .
the quasiparticle spectral weight , concentrated at @xmath85 in the free case , is now spilled into the region @xmath102 ( fig . [ weight ] ) :
the density of states exhibits a pseudogap ( fig .
[ pseudogap ] ) , somewhat more pronounced than it was found in the @xmath0-matrix approximation : it now vanishes as @xmath104 near the fermi level .
the results of this section can be understood in a simple intuitive way .
even though we have suppressed spatial and temporal fluctuations of the pairing field , its strength does not yet have a definite value ; the fermion pole , instead of being shifted to a definite new position , is smeared into a cut along some part of the real axis . nevertheless , the existence of a scale @xmath105 in the field distribution results in the fermion spectral function reminiscent of the bcs one . the @xmath0-matrix equation for the fermion self - energy ( [ slow sigma ] ) , which includes a _ dressed _ fermion propagator , is equivalent to partial summation of the diagrams in sec .
[ frozen fluctuations ] : only those without intersecting dashed lines are included .
for instance , diagram ( b ) in fig .
[ gauss ] is present , while ( c ) is left out .
the number of diagrams in the @xmath26-th order satisfies the recursion relation @xmath106 ( @xmath107 by definition ) with the solution @xmath108 which grows as @xmath109 for a large @xmath26 .
therefore , the series @xmath110 has the radius of convergence equal to 1/4 and the green s function @xmath111 is analytical ( hence real ) when @xmath112 .
the spectral weight is contained in the finite region @xmath113 .
the above discussion shows that the @xmath0-matrix approximation includes , at least partially , effects of quantum fluctuations of the pairing field on the fermion propagator .
this is achieved through the use of a dressed fermion propagator in the equation for the fermion self - energy .
the number of terms in the resulting perturbation series grows quickly enough to make the result non - analytical in a larger region than in the bcs broken phase .
inclusion of a greater number of terms ( the full series ) only slightly enhances this effect .
this should be contrasted to the situation when a _ bare _ fermion propagator is used in the fermion self - energy ( [ slow sigma ] ) .
since there is only one graph in each order of the perturbation series for the propagator @xmath29 , such a treatment essentially amounts to throwing away the quantum nature of cooper pairs .
no broadening is then found in the limit of slow fluctuations .
with much of the spectral weight removed from the vicinity of the fermi surface , low - energy bosonic excitations may represent a propagating mode . in this respect , such a normal state
resembles a bcs superconductor just below @xmath1 , where the opening of a true gap inhibits the decay of pairs.@xcite therefore , a reasonable estimate of the boson energy spectrum can be obtained from eq .
[ pi(p ) ] by using the dressed fermion propagator in the bcs form ( [ bcs g ] ) , which constitutes the two - pole _ ansatz _ of ref .
[ micnas ] .
inclusion of fermion lifetime effects makes cooper pairs unstable but is not expected to produce qualitative changes in the energy spectrum.for a short - range instanteneous attraction , @xmath114 , @xmath115 with @xmath116 at this level of approximation ( no boson decay because of a full fermion gap ) , @xmath117 , analytically continued from matsubara frequencies to the rest of the @xmath118 plane , can be expanded in powers of @xmath118 and @xmath119 : @xmath120 or @xmath121 the mass term appears because the thouless criterion no longer applies : @xmath122 in the normal state .
the existence of _ two _ cooper - pair poles at a given momentum @xmath119 , at the frequencies @xmath123 , reflects the fact that correlated propagation is possible not only for two fermionic particles , but also for two holes . if a pair is made out of fermionic excitations just around the fermi surface , where the density of states is the same for particles and holes , the residues of the two poles are equal ( up to a sign )
. a varying density of states will make the two poles less symmetric , both in terms of their residues and positions , but still low - energy cooper pairs will have two poles
. this trend can be clearly seen in the numerically obtained two - particle density of states plotted in fig . 4 of ref . [ micnas ] .
low - momentum pairs have two somewhat asymmetric peaks dispersing with momentum .
the asymmetry is related to the instanteneous nature of the interaction in the attractive hubbard model : the number of particle states involved in the formation of a pair greatly exceeds that of hole states at low fermion densities .
no holes contribute to the formation of pairs with higher momenta , so that there is no pole at negative frequencies in this case .
such pairs , however , represent fast fluctuations of the pairing field and are irrelevant to the formation of the pseudogap .
the numerical data also show an increasing lifetime of cooper pairs as @xmath124 , reflecting a stronger depletion of the fermion density of states near the fermi energy . when fermionic excitations across the gap are frozen out , the cooper - pair propagator at zero momentum is @xmath125 where @xmath126 . with the exception of the zeroth - order term @xmath127 ,
coefficients of the taylor series ( in powers of @xmath118 ) are insensitive to large-@xmath128 behavior of the density of states , and are thus determined by the lower energy scale , _
i.e. , _ the gap width @xmath129 , provided that variations of the bare density of states are small on that energy scale . in this weak - coupling limit
, the terms odd in @xmath118 vanish because @xmath130 is an odd function near @xmath131 .
quartic and higher - order terms can be dropped as long as @xmath132 .
then @xmath133 where @xmath134 is of order @xmath135 .
more exactly , @xmath136 expansion in powers of @xmath119 allows one to determine the characteristic speed of bosons @xmath137 , which is expected to be smaller than the fermi velocity @xmath9 on physical grounds . in the weak - coupling limit ,
we recover the result of bogoliubov , @xmath138 in @xmath139 dimensions .
this is roughly consistent with the dispersion of cooper - pair poles at low momenta inferred from the data of ref .
[ micnas ] . at a given temperature , two as yet undetermined low - energy parameters of the boson field the mass @xmath140 and the pole residue @xmath141 control the intensity of fermion scattering by the pairing field and thus determine the ( also unknown ) fermion energy gap , as discussed in some detail below .
in addition , eq . [ z ] relates @xmath141 to the energy gap . using this information
, it should be possible to determine two out of the three unknown parameters . in the following section
, the mass @xmath140 is found as a function of temperature .
we now return to the analysis of the fermion self - energy ( [ sigma bosonized ] ) using the approximate boson propagator ( [ pseudorelativistic ] ) .
the following expression for the fermion self - energy results : @xmath142 the sum over the bosonic matsubara frequencies @xmath118 has been converted into an integral around the poles of @xmath143 ; @xmath144 is the boson energy .
it is convenient to write the fermion propagator in terms of its spectral weight : @xmath145 after deforming the integration contour in the standard way , one finds the self - energy as a sum of two terms , which correspond to emission ( @xmath146 ) and absorption ( @xmath147 ) of a cooper pair : @xmath148 the part proportional to the boson occupation number @xmath149 can be regarded as induced emission or absorption with the rate proportional to the spectral intensity of the pairing field @xmath150 .
the fermionic term @xmath151 is responsible for `` spontaneous '' processes .
if the pseudogap formation is governed by low - energy modes , the induced term is dominant since @xmath152 at low energies .
the rate of spontaneous emission and absorption of pairs is estimated in the appendix ; it can be neglected in @xmath153 dimensions . upon shifting the integration variable @xmath154 ,
@xmath155 bosons are restricted to low - energy modes with @xmath156 of order or less than @xmath157 . in the case
when the fermion spectral weight is distributed over a wider interval of energies , the variation of the fermion spectral weight on the energy scale of @xmath157 can be neglected .
for the same reason , we can neglect the variation of the fermion momentum in the integrand , provided that fermion and boson velocities are of the same order ( recall that @xmath138 ) .
the integration over @xmath158 yields the fermion propagator @xmath159 and the self - energy has the form inferred previously ( [ slow sigma ] ) : @xmath160 where the average fluctuation ( intensity ) of the pairing field is @xmath161 essentially the same result can be obtained by considering a classical field @xmath162 with the lagrangian density determined by ( [ pseudorelativistic ] ) , @xmath163,\ ] ] in thermal equilibrium .
application of the equipartition theorem yields @xmath164 which is the classical analogue of eq . [ fluctuation of delta ] . by substituting the value of the boson residue ( [ z ] ) into eq .
[ fluctuation of delta ] , we obtain the self - consistency condition mentioned above .
remarkably , the width of the pseudogap cancels out and the resulting equation implicitly determines the boson mass as a function of temperature : @xmath165 where @xmath166 , @xmath167 is the boson occupation number , and @xmath66 is the density of bare fermion states .
it is immediately obvious that there is no bose condensation in @xmath153 dimensions : for @xmath168 , the integral diverges at the lower limit . in @xmath169 dimensions ,
this equation predicts a condensation temperature @xmath170 of order @xmath13 .
this result should not be taken at face value because it was derived for a weakly attractive degenerate fermion gas .
( also , as discussed in the appendix , the absence of a low energy scale implies that pairing fluctuations are not slow . ) nevertheless , the existence of such a high temperature scale is justified if one considers the limits of weak and strong attraction between fermions . in the weak - coupling limit , there are no pairs at the fermi temperature , they form at a much lower temperature @xmath171 ; therefore , the long - overdue condensation occurs immediately , which explains why the mean - field bcs approach works so well . in the opposite limit of local pairs (
turned into hard - core bosons ) , condensation indeed occurs when the gas becomes degenerate .
next , we estimate the condensation temperature in @xmath172 dimensions , which may be relevant to highly anisotropic cuprate superconductors . recalling the asymptotic behavior of the riemann zeta - function @xmath173 we obtain @xmath174 when @xmath175 . for comparison ,
the condensation temperature of a bose gas made of very small fermionic pairs is only higher by a factor @xmath176 .
( in both cases , @xmath13 is the fermi energy of the ideal fermion gas with the same mass and density of particles.@xcite ) it is thus clear that the condensation of cooper pairs in the presence of a pseudogap occurs well below the fermion degeneracy temperature and a clear separation of energy scales ( [ separate scales ] ) is possible for moderately weak attraction strengths .
if the attraction is strong , we end up with tightly bound pairs , in the local boson limit . in the case of a very weak attraction , the bcs pair formation temperature @xmath6 is lower than @xmath1 of eq .
[ t_c near d=2 ] ; the bcs model takes over .
i have considered the interaction between fermions and gaussian pairing fluctuations ( cooper pairs without self - interaction ) .
a well - defined pseudogap regime is found in @xmath172 spatial dimensions , when the condensation temperature is much lower than the bcs mean - field one .
the quantum character of cooper pairs , related to the unobservable nature of the pair wavefunction @xmath71 , leads to considerable broadening of the bogoliubov quasiparticle peaks near the fermi surface .
even when non - uniform configurations of the field @xmath71 are suppressed at low temperatures , fluctuations of its _ amplitude _ blur the strength of the pairing field , making the position of the quasiparticle poles at @xmath177 uncertain .
a well - pronounced suppression of the fermion density of states near zero energy makes room for long - lived pair states inside the gap . at low momenta and frequencies ,
they represent a bogoliubov sound - like mode with a non - zero mass .
the mass is determined from a self - consistency condition that links the width of the pseudogap to the mean fluctuation of the pairing amplitude .
zero mass signals the onset of bose condensation . in @xmath172 dimensions , the condensation temperature scales as @xmath178 , which is @xmath179 times the condensation temperature for the corresponding ideal bose gas .
this may explain the doping dependence of @xmath1 in underdoped cuprates .
approximations made in this work are twofold .
first , an infinite hierarchy of equations for @xmath26-particle green s functions is replaced with the self - consistent @xmath0-matrix approach of brueckner , which partially accounts for gaussian pairing fluctuations . by solving an exact model with an independent pairing field ,
i show that the self - consistent @xmath0-matrix approximation takes into account the amplitude fluctuations of cooper pairs in the normal state , a feature missing in other approximate models.@xcite further approximations made in this paper are necessary to obtain an analytic solution .
firstly , thermally excited non - uniform configurations of the pairing field have little effect on the fermion propagator , provided that a clear separation of energy scales , @xmath180 , exists .
this appears to be the case in @xmath172 dimensions , _
i.e. , _ for highly anisotropic , almost two - dimensional systems . secondly , a residual density of fermionic states in the pseudogap , leading to the decay of cooper pairs , is assumed to have negligible influence on the pair energy spectrum .
a numerical work using the self - consistent @xmath0-matrix approach@xcite seems to indicate that this is a reasonable assumption : low - energy cooper pairs represent a propagating mode with two cooper - pair poles dispersing with the pair momentum . in order to ascertain that propagating cooper pairs in the pseudogap are not an artifact of the self - consistent @xmath0-matrix approximation ,
it is necessary to look for them in quantum monte - carlo simulations on two - dimensional lattices .
there are many questions that remain open .
for instance , in the case of a superconducting order parameter with nodes , the density of fermion states is less strongly suppressed .
will cooper pairs be stable enough to represent a propagating mode ?
also , what causes the strong broadening of fermion quasiparticle peaks above the temperature at which pairs are formed ?
the idea of this work has originated in conversations with y. j. uemura .
i thank a. s. blaer , d. m. eagles , t. d. lee , r. micnas , m. randeria , m. v. sadovskii , and especially h. c. ren for helpful discussions .
this work has been supported in part by the u. s. department of energy ( grant de - fg02 - 92 er40699 ) and by nedo ( japan ) .
spontaneous emission or absorption of cooper pairs is not restricted to low boson momenta and thus represents the effect of fast fluctuations .
the emission or absorption rate depends on the phase space available to the products of a fermion decay ( a pair and a hole ) , which tends to be smaller in a lower number of dimensions @xmath139 .
this indeed is confirmed by a simple calculation below . in @xmath169 dimensions ,
the imaginary part of @xmath181 is linear in @xmath182 and thus leads to a renormalization of the quasiparticle spectral weight . in @xmath153 dimensions ,
a constant imaginary part is introduced , which merely broadens the quasiparticle peak .
besides , the existence of a small parameter in the pseudogap regime ( the ratio @xmath183 ) makes this additional broadening insignificant in @xmath184 dimensions .
let us estimate the impact of `` vacuum '' fluctuations on the self - energy of a fermion with momentum @xmath185 and bare energy @xmath186 .
inasmuch a hole created in the decay can be far away from the fermi surface , we can neglect the influence of pair scattering on its spectrum and treate the hole as a stable particle with momentum @xmath187 and energy @xmath188 neglecting the curvature of the energy surface @xmath59 ; @xmath189 is the angle between @xmath185 and @xmath119 .
the boson energy is taken to be @xmath190 .
then @xmath191 energy conservation requires that @xmath192 for definiteness , let us assume that the frequency of the incoming fermion @xmath193 .
then a pair can be emitted ( @xmath194 ) when @xmath195 and absorbed ( @xmath196 ) when @xmath197 .
pair emission is accompanied by absorption of a second fermion with the energy in the range @xmath198 .
the second fermion can , thus , be taken from _ below _ the fermi surface and
the process will take place even at low temperatures ! on the contrary ,
absorption of a pair requires emission of a second fermion with a negative energy @xmath199 . at low temperatures ,
such states are occupied and pair absorption is suppressed . in @xmath169 dimensions ,
the element of phase space reduces to @xmath200 .
the integration over the boson energy @xmath201 is elementary because a @xmath202-function is present : @xmath203,\ ] ] the first term in brackets ( emission of a pair ) grows linearly with @xmath182 . by inserting @xmath141 from ( [ z ] ) , we obtain @xmath204 as @xmath205 .
thus , spontaneous pair emission leads to a significant renormalization of the fermion propagator ( the ratio @xmath206 is not small for the pseudogap regime in @xmath169 ) and can not be neglected . in @xmath153 dimensions ,
pair emission has a smaller available phase space : @xmath207 the self - energy does not vary with @xmath182 if @xmath208 is replaced with the step - function .
this channel of decay is characterized by the rate @xmath209 where @xmath210 . in @xmath172 dimensions , @xmath211 and
this additional linewidth is small in comparison with the broadening caused by induced fluctuations , of order @xmath212 .
therefore , the effect of fast spontaneous fluctuations can be neglected in @xmath172 dimensions .
[ uemura plot ] y. j. uemura _ et al .
. lett . * 62 * , 2317 ( 1989 ) .
see also y. j. uemura , in _ proceedings of the workshop on high-@xmath1 superconductivity and the @xmath213 family , beijing , 1994 , _ edited by s. fung and h. c. ren ( gordon and breach , 1994 ) .
a. schmid , z. phys .
, * 231 * , 324 ( 1970 ) .
the main difference between the present work and schmid s lies in the properties of pairs : he considers a _ diffusing _ pairing field , as opposed to a propagating one .
i thank r. micnas for pointing out this reference to me .
[ dissipation ] r. v. carlson and a. m. goldman , phys .
lett . * 31 * , 880 ( 1973 ) ; _ ibid .
, _ * 34 * , 11 ( 1975 ) .
these authors stress that , as soon as a superconducting gap opens up below @xmath1 , pairing fluctuations become a _ propagating _ , sound - like mode .
well below @xmath1 , this mode is pushed up to the plasmon frequency . | i present a simple analytical model describing the normal state of a superconductor with a pseudogap in the density of states , such as in underdoped cuprates . in nearly two - dimensional systems , where the superconducting transition temperature is reduced from the mean - field bcs value , cooper pairs may be present as slow fluctuations of the bcs pairing field . using the self - consistent @xmath0-matrix ( fluctuation exchange ) approach
i find that the fermion spectral weight exhibits two bcs - like peaks , broadened by fluctuations of the pairing field amplitude .
the density of states becomes suppressed near the fermi energy , which allows for long - lived low - energy cooper pairs that propagate as a sound - like mode with a mass .
a self - consistency requirement , linking the width of the pseudogap to the intensity of the pairing field , determines the pair condensation temperature . in nearly two - dimensional systems , it is proportional to the degeneracy temperature of the fermions , with a small prefactor that vanishes in two dimensions . |
recent developments that lead to extending of plasmonics to the mid- and far - infrared ( terahertz ) ranges of frequencies constitute an important direction of modern photonics @xcite .
surface plasmon polaritons ( spps ) are partially longitudinal @xmath0-polarized evanescent electromagnetic waves that propagate along a metal - air interface as a result of collective electron oscillations coupled to an external electromagnetic field @xcite . in practical applications
it is important to understand how spps propagate along a curved surface @xcite .
for this reason , the theory of spps on a cylinder attracted a lot of attention in the past and recent years @xcite .
a dispersion equation for the transverse magnetic ( tm ) modes was derived by m. berry in an artificial case of lossless metal @xcite and later by k. hasegawa et al . in a general case @xcite .
a similar equation appears in the theory of diffraction of a wave on a transparent cylinder @xcite .
it involves the bessel and hankel functions of a large argument and a large complex order which makes its numerical evaluation for practical problems difficult .
different approaches have been developed to overcome this difficulty as reviewed in ref .
@xcite but none of them was especially successful .
the standard approach to the derivation of spps @xcite assumes that the radial distribution of the magnetic field of a plasmon near the surface of a metallic cylinder is described by the hankel function of the first kind , @xmath1 , with @xmath2 , @xmath3 the frequency of the mode , and @xmath4 the radius in cylindrical system of coordinates .
the hankel function of the second kind , @xmath5 , is explicitly or implicitly discarded .
this approach is usually justified by the fact that at large distances , @xmath6 , the asymptotic representation of the hankel function @xmath7 describes a radial distribution of a wave propagating away from the cylinder , whereas @xmath5 stands for a wave , propagating to the cylinder .
the situations however becomes more complicated if one takes into account that , in the case of finite conductivity of the metal , the index @xmath8 , which has to be found from a dispersion equation , is a large complex number . in this case
, @xmath7 can experience a large exponential growth in the radial direction in the vicinity of the metal surface , meaning that the mode is not actually localized near the surface . in our opinion , an adequate approach to the study of spps
is to analyze their spatial ( radial plus azimuthal ) distribution in the vicinity of the metal surface , @xmath9 , where @xmath10 is the cylinder radius ( see fig .
[ fig : geometry ] below ) .
we introduce a small parameter @xmath11 , where @xmath12 is a transverse localization length of the mode and derive equations in which small terms in parameter @xmath13 are systematically neglected .
a similar method was previously successfully used in the theory of synchrotron radiation in a curved waveguide @xcite .
here we use it for investigation of the surface electromagnetic waves ( sw ) on a metal - air interface with given and fixed radius of curvature . solving these equations
we obtain expressions for the fields in terms of the airy - fock functions , avoiding in this way the numerical difficulties arising when one deals with the hankel functions of large argument and large order . in this our study
we find that , indeed , the localized solutions in the form of @xmath7 used in @xcite and some subsequent papers ( see eg .
@xcite ) may exist only in a narrow range of values of the dimensionless surface impedance @xmath14 . on the contrary
, we find that sws which are expressed through the hankel functions of the second kind @xmath15 , do exist for any @xmath14 provided that the radius of the cylinder is not too small .
we named these two types of surface waves correspondingly the outgoing and incoming surface waves .
we also note that in a planar geometry an ordinary surface wave belongs to the type of incoming surface waves . as to the outgoing modes discussed in refs .
@xcite , they can be identified as creeping waves ( cw ) .
in contrast to the surface waves , the amplitude of cws grows in radial direction in the vicinity of the metal ; in ref . @xcite these waves were characterized as wave - guide modes . the term `` creeping waves '' is used in the theory of diffraction @xcite for the refracted waves over convex conducting or dielectric surfaces where an extra attenuation due to the curvature is present .
the foundations of the theory were laid out in earlier works of v.a .
fock @xcite who considered propagation of radio waves around the earth . as explained in ref .
* chapt . 5 )
, the creeping waves rapidly damp in the shadow region created by a convex obstacle on the way of a wave propagation .
they damp both in the direction of propagation and towards to the surface of the obstacle . in other words ,
the amplitude of a creeping wave grows with the distance along the normal to the surface from the shadow region to the lit one . on the contrary
, the amplitude of the surface waves decays with the distance from the surface .
this qualitative difference of sws from cws , which are also known as `` whatson modes '' , is discussed in ref .
@xcite .
the paper is organized as follows .
a paraxial equation for electromagnetic waves near a curved surface is derived in sec . [ s2 ] .
it is used in sec .
[ s3 ] to obtain dispersion equations of various surface electromagnetic waves on a curved surface ; we introduce here the notion of outgoing and incoming surface waves . a numerical solution to these equations
is discussed in sec .
analytical solutions are presented in sec .
all these sections deal with a convex geometry , where the metallic medium is a cylinder surrounded by air .
a concave case is considered in sec .
[ s6 ] , where it is assumed that a hollow conducting cylinder is filled by air . in sec .
[ s7 ] we show how the airy - fock functions are matched to the hankel functions in a relevant range of parameters thus providing a bridge between our results and previous studies in ref
. @xcite . finally , in sec .
[ s8 ] we summarize main results .
in this section we will derive equations for the electric and magnetic fields near the surface of a conducting cylinder .
the geometry of the problem and the choice of the coordinate system is shown in fig .
[ fig : geometry ] .
we use the cylindrical coordinate system @xmath4 , @xmath16 , @xmath17 and the notation @xmath18 for the difference @xmath19 .
the metal surface is located at @xmath20 . the cylinder radius .
a conducting medium occupies either an interior ( convex geometry ) or exterior ( concave geometry ) of the cylinder .
the radial coordinate @xmath4 is measured from the axis of the cylinder , and @xmath19 in counted from the cylinder surface . a transverse magnetic ( tm )
wave under consideration has magnetic field @xmath21 directed along the axis @xmath17 of the cylinder . ]
we first assume that the metal occupies the region @xmath22 ; the region @xmath23 has the dielectric permeability @xmath24 .
this is so called _ convex case _
@xcite , where a metallic circular cylinder is surrounded by an empty space ( or by air ) .
later on , we will extend our results to the _ concave case _ by assigning an opposite sign to the radius @xmath10 .
as to the current and subsequent sections , except for sec .
[ s6 ] , we suppose that @xmath25 .
we assume that all components of the field have the following dependence on time and the azimuthal angle @xmath16 @xmath26 with amplitude @xmath27 independent of the coordinate @xmath17 .
moreover , we restrict ourselves to the case of transverse magnetic ( tm ) waves although it has been recently reported @xcite that transverse electric ( te ) surface waves may exist on a cylindrical metal - air interface in contrast to the absence of such waves in a planar geometry ( see eg .
hence , we assume that the magnetic field @xmath28 has only the @xmath17-component , and the electric field @xmath29 has both the @xmath4- and @xmath16-components . to keep track of order of magnitudes
, we will assume that @xmath30 is of the order of @xmath31 and the radius @xmath10 is of order of @xmath32 , where @xmath33 is a formal small parameter of the problem .
the wavenumber @xmath34 and the frequency @xmath35 are assigned the order of @xmath36 .
we will clarify our ordering scheme in sec .
[ s3 ] in more details . from the radial component @xmath37 of the maxwell equation @xmath38 ( here and below we use the gaussian system of units )
, we find to the lowest order @xmath39 this is the relation that holds in a plane electromagnetic wave ; in our case it is satisfied approximately to order @xmath40 . from the azimuthal component of the same equation @xmath41 we express @xmath42 in terms of @xmath43 , @xmath44 as to the function @xmath45
, it obeys the equation @xmath46substituting @xmath47 and expanding in small ratio @xmath48 we find @xmath49where @xmath50 equation describes electromagnetic field in the air near the cylinder surface propagating mainly along the axis @xmath16 ; for this reason it will be referred to as the _ paraxial equation _ by analogy with the paraxial approximation in geometrical optics .
eq . should be complemented by a boundary condition on the surface of the metal @xmath20 . assuming the skin depth in the metal much smaller than the reduced wavelength @xmath51 , we use the leontovich boundary conditions @xcite @xmath52 \end{gathered}\ ] ]
where @xmath53 is a unit vector normal to the boundary and directed from metal to air , and @xmath14 is a dimensionless surface impedance .
the parameter @xmath14 is given by the following formula ( see , e.g. , @xcite ) @xmath54 where @xmath55 is the metal conductivity .
good conductors with @xmath56 are characterized by a low surface impedance , so we assume that @xmath57 and assign @xmath14 the order of @xmath13 .
the condition implies for the dielectric permittivity of the metal @xmath58 to be large , i.e. @xmath59 in this case a general dispersion relation for sw on a planar metal - air interface @xcite @xmath60 is reduced to @xmath61 it assumes for the parameter @xmath62 to be of order of @xmath63 .
it can be seen then from eq . that @xmath48 should also be of order of @xmath63 , as assumed above , for the curvature effect to modify substantially the dispersion relation of the surface waves on a metallic cylinder .
to avoid possible misunderstanding , we note that the condition implies that the wave frequency is lower than the plasma frequency , @xmath64 .
thus , a high frequency branch of the surface wave with @xmath65 , which has an electrostatic limit @xmath66 at @xmath67 , is outside of the scope of our analysis . as to the low frequency branch , @xmath68 ,
it is typical for the experiments in the terahertz range of frequencies as at the novosibirsk free electron laser facility @xcite .
practical estimations of the surface impedance for these experiments can be found in ref .
some papers ( see eg .
@xcite ) assumed the drude expression @xmath69 for the dielectric permittivity instead of .
although the drude approximation is not applicable to realistic metals in the frequency range of interest , we will be able to compare our results against these papers in the limit @xmath70 .
it can be deduced from the standard theory of surface electromagnetic waves @xcite that in a planar geometry a low frequency branch of sw evanesces inside the conductor ( @xmath22 ) as @xmath71 with @xmath72 which means that using the leontovich boundary condition is justified if @xmath73 outside of the metal ( @xmath23 ) , in the planar geometry @xmath74 with @xmath75 which means that sw is formed if @xmath14 is not purely real . looking for a solution of a boundary value problem in the form of surface wave with a given real positive frequency @xmath76 , we see that the wavenumber @xmath34 in eq
. should have a positive imaginary part @xmath77 and the parameter @xmath78 should have a positive real part @xmath79 any solution that obeys the inequality describes a wave propagating and simultaneously damping in the positive direction of the azimuthal angle @xmath16 , and eq .
ensures that sw is localized near the conductor surface .
these two conditions mean that physically feasible values of @xmath14 obey the conditions @xmath80 as can be proved independently from a consideration of the energy dissipation inside a conducting body @xcite .
in what follows we will refer to this property of the surface impedance as @xmath81 for what follows , it is important to examine at this point the radial dependence of a planar sw in combination with its temporal dependency : @xmath82 .
\end{gathered}\ ] ] putting here explicit expression for @xmath78 yields @xmath83 .
\end{gathered}\ ] ] the last term in the square brackets shows that the field is spatially localized near the metal - air interface at @xmath20 , whereas the first term describes a wave propagating _ towards _ the interface from the side of large @xmath18
. we will refer to such sws as _ incoming _ surface waves .
note that all previous treatments of sws on a curved metal - air interface were focused on the _ outgoing _ surface waves that are absent in the planar geometry .
returning to cylindrical geometry the boundary condition reads @xmath84 which using reduces to @xmath85 to proceed further we denote @xmath86 and introduce the dimensionless coordinate @xmath87 and dimensionless parameter @xmath88 this casts the paraxial equation into the form @xmath89 and the boundary condition at @xmath90 reads @xmath91 where @xmath92 we also formulate below an additional boundary condition when @xmath93 . note that in our ordering scheme all the dimensionless parameters @xmath94 , @xmath95 and @xmath96 have zero order @xmath31
. general solution of eq .
( [ 4:3 ] ) involves the airy functions
@xmath97 and @xmath98 , @xmath99 where @xmath100 are unknown constants . using asymptotic formulas for large negative values of the argument (
9.7.9 and 9.7.11 ) , @xmath101
we see that each function contains both outgoing ( propagating outward from the cylinder ) and incoming ( propagating inward from infinity ) waves .
we first consider the case with an outgoing wave .
physically it means that a surface wave radiates and its energy is leaking radially to infinity .
this same assumption is adopted in refs .
it introduces an extra damping to the wave , which is launched by a source in direction tangent to the surface of the cylinder .
it is easy to see that asymptotically @xmath102 which ( with assumed time dependence @xmath103 ) corresponds to the radial propagation from the cylinder surface to infinity ; note that eqs . and
are written for the range @xmath104 .
this combination of the airy functions @xmath105 \end{gathered}\ ] ] we will refer to as a solution of the first kind since it can also be expressed through the airy - fock function @xmath106 of the first kind ( see below ) .
the boundary condition for this wave reads @xmath107 it represents a dispersion equation to be resolved regarding @xmath95 at a given parameter @xmath96 . the condition eq .
restricts a goal range for @xmath95 to upper half of the complex plane , ie .
@xmath108 and an accurate treatment of the wave at @xmath109 should account for complex nature of @xmath110 .
this can be accomplished by expanding the asymptote for @xmath111 to the following form @xmath112 where the factor @xmath113 exponentially grows as long as the condition eq
. holds .
it means that the solution of the first kind given by eq .
can hardly describe a surface wave .
it would not be surprising since , as we have seen , a planar sw is of the incoming type . on the contrary ,
a solution of the second kind , @xmath114 , \end{gathered}\ ] ] describes a radially incoming wave with radially decreasing amplitude at @xmath109 , where @xmath115 and the factor @xmath116 exponentially vanishes .
a corresponding dispersion equation reads @xmath117 the combinations of the airy functions @xmath97 and @xmath98 in eqs . and can be recognized as the airy - fock functions @xcite @xmath118 it can be checked that @xmath119 earlier investigation of eq .
can be found in @xcite , whereas eq . seems to escape attention of the researchers .
[ fig : airycomplexmap ] illustrates the results of numerical search of complex roots of the dispersion equation for the wave of the first kind . using wolfram mathematica @xcite to calculate the airy functions , the roots of eq . can be found numerically .
they are located on the intersections of the zeroth isolevels of the real and imaginary parts of lhs of eq .
( multiplied by the denominator of the first term ) .
these isolevels are represented in figs .
[ fig : airycomplexmap ] by the blue ( solid ) and purple ( dashed ) lines , respectively .
inspecting these figures shows that eq .
has many roots .
in particular , a ( presumably ) infinite series of roots is located near the ray @xmath120 .
we found that the roots marked by the white circles in fig .
[ fig : airycomplexmap ] have radial profiles of @xmath121 monotonically growing with @xmath94 as expected for the creeping waves ( see discussion of cw in the introduction ) .
hence we associate these roots with the creeping waves .
the only exception from monotonically growing modes is represented by the root located at the center of the white square in fig .
[ fig : airycomplexmap]a , b , and c. the radial profiles of the modes , corresponding to this root are plotted in fig . [ fig : airyradialprofile]a , b , and c , respectively for 3 different values of the parameter @xmath96 . comparing these plots
, we see that with increasing @xmath96 the radial profile of these modes is gradually converted to a radially decaying pattern , which is expected for a surface mode localized in the vicinity of a metal - air interface . such a localized mode we identify with a surface electromagnetic wave ( sw ) .
it is important to note , that no sws were found if @xmath122 . in the case , where @xmath123 , a single surface - like mode is formed near the metal - air interface for every given value of @xmath96 if @xmath124 . for a small @xmath96 ,
an amplitude of such a mode initially diminishes but eventually becomes to grow at larger @xmath94 as shown in fig .
[ fig : airyradialprofile],a .
we note also , that the sw root is visibly well separated from the other roots in fig .
[ fig : airycomplexmap ] when sw becomes radially localized ( as in fig . [ fig : airycomplexmap],b , c ) , whereas it is located in the vicinity of the other roots in the case if sw is not well localized ( as in fig .
[ fig : airycomplexmap],a ) .
a similar behavior of the sw roots was noted in ref .
it was stated there that the sw roots freely migrate in the complex map of the parameter @xmath125 , whereas the cw roots remain tightly confined to the immediate vicinity of the stokes line of the hankel functions . the radial energy flux given by the radial component @xmath126 of the poynting vector for the sw
is shown in fig .
[ fig : airypoynting ] .
it is directed inward to the metal - air interface in its vicinity but is reversed outward at larger distances @xmath127 , where it might be too small as in fig .
[ fig : airypoynting],c .
the change of sign of @xmath128 is a manifestation of the two - dimensional nature of the wave which propagates mainly along the metal surface but also damps in that direction because of the energy leakage both into the metal and radially to the infinity . as to the wave of the second kind , eq .
has a sw - like solution for the entire allowed range @xmath129 provided that @xmath130 .
the results of numerical computations are shown in figs .
[ fig : airycomplexmapplus ] , [ fig : airyradialprofileplus ] , [ fig : airypoyntingplus ] . serial roots , which might also be thought of as creeping waves , are located near the ray @xmath131 as shown fig .
[ fig : airycomplexmapplus ] .
note however that @xmath132 for these roots and , hence , corresponding modes amplify in the direction of propagation along the metal surface due to energy influx from large radii .
alone root , which corresponds to sw , is well separated from the cw roots .
it passes to the upper half of the complex plane @xmath133 if @xmath124 but is located in the lower half - plane @xmath134 if @xmath135 as shown in fig .
[ fig : airycomplexmapplus],a . in the latter case , the wave amplitude eventually grows with @xmath94 as in fig .
[ fig : airyradialprofileplus],a . in case of @xmath124
the radial profile of @xmath136 monotonically decreases to zero as shown in figs .
[ fig : airyradialprofileplus],b , c .
the energy radial flux @xmath128 is everywhere directed inwards as shown in fig .
[ fig : airypoyntingplus ] .
this fact distinquishes the surface wave of second type from the first one .
both types are characterized by inward radial energy flux in the vicinity of the metal - air interface as could be expected for a realistic metallic cylinder that should dissipate the wave energy , however at large distances the radial energy flux in the wave of the first type reverses its direction .
for this reason we call the surface wave of the first type also the outgoing wave whereas the surface wave of the second type is referred to as the incoming wave .
we note also that the radial profile of cws ( not shown in fig .
[ fig : airyradialprofileplus ] ) exhibits growth in the vicinity of the metal - air interface in contrast to the surface waves .
we found that for any value of @xmath96 there is at most one surface mode of both types .
let us see how we can recover a sw on a planar surface in the limit @xmath137 .
for a large @xmath10 the parameter @xmath96 also becomes large , @xmath138 , and as we will see from the result , @xmath139 .
we start from eq . for the wave of first kind and use the asymptotic expansion ( * ?
9.7.5 and 9.7.13 ) @xmath140 valid in the range @xmath141 and @xmath139 .
this reduces eq . to @xmath142
squaring this equation and using and we obtain @xmath143 which can readily be recognized as the dispersion relation of a planar surface wave ( psw ) . having obtained the solution we need to verify if the argument of @xmath95 lies within the intersection of the range that was assumed in the asymptotic expansion and the target range .
it follows from that @xmath144 and one finds that this is indeed the case if @xmath145 ( here we take into account that the argument of @xmath14 lies in the interval given by ) . radial profile of psw near the metal - air interface can be recovered using the asymptotic expression and expanding the function for @xmath146 : @xmath147 the condition @xmath146 means that the surface wave is planar at the distance @xmath148 from the metal - air interface . at larger distances , @xmath149 ,
it is transformed into an outgoing wave with exponentially growing amplitude as is seen in figs [ fig : airyradialprofile]a - c . to obtain a radiative correction to the dispersion relation of psw
, one needs to take a more accurate asymptotic expression for the airy functions . although @xmath150 and @xmath151 are entire functions analytical in the entire complex plane of their argument without any cuts , they have different asymptotic expressions in various segments of the complex plane which are separated by stokes lines @xmath152 and @xmath153 . in particular , as it has been already mentioned , the asymptotic expression eq .
is valid if @xmath154 . recall that we are looking for a solution with @xmath110 in the range @xmath155 which describes the waves propagating and damping in the positive direction of the angle coordinate @xmath156 .
an appropriate asymptote for this case ( * ? ? ?
9.7.5 and 9.7.13 ) @xmath157 contains an additional term as compared to eq . and is appropriate for the range @xmath158 .
the second term in is negligible ( subdominant ) if @xmath159 but it dominates if @xmath160 exceeds @xmath161 . putting into eq .
casts it to the form @xmath162 if @xmath163 and @xmath164 , the left - hand side ( lhs ) of this equation is large , which means that the denominator of the right - hand side ( rhs ) tends to zero , thus reproducing eq . .
we now take @xmath165 from eq . as a first approximation and put it back into eq . everywhere except for the denominator in rhs .
this yields a corrected expression @xmath166 , \end{gathered}\ ] ] where the second term stands for the radiative addition to the outgoing sw dispersion relation .
it is responsible for additional damping of sw , which is characterized by @xmath167 .
in particular , a planar surface wave does not damp in the direction of propagation in case of pure imaginary surface impedance , @xmath168 , since the first term in turns out to be purely real .
however , the second term is imaginary giving rise to radiative damping of the surface wave . according to
, this sort of damping becomes exponentially weak as the curvature radius @xmath10 tends to infinity .
this result was earlier obtained for a particular case of purely imaginary ( inductive ) surface impedance by m. berry @xcite ( with tow times smaller coefficient before exponential term in his eq .
28 ) and by n. logan and k. yee in ref .
@xcite ( see their unnumbered equation after eq .
the effect of the curvature of the metal - air interface on the wave of the first kind becomes much stronger if @xmath14 acquires a real part .
depending on the value of @xmath169 the second term in eq . can be either decreasing or growing function of @xmath10 .
when @xmath169 approaches or exceeds the value @xmath170 , the second term in eq . grows exponentially with @xmath10 and for large values of @xmath10 exceeds the first one . since we used a perturbation theory , which assumes that the second term should be small in comparison with the first one , eq .
becomes invalid in this case .
our numerical calculations indicate that eq .
has no solution in the form of sw if @xmath171 thus , an outgoing sw does not exist for all physically reliable values of the surface impedance @xmath14 even in the limit of large radius @xmath10 of conducting cylinder
. numerical study of the problem detailed in sec .
[ s4 ] showed that eq . has a large ( perhaps , infinite ) number of solutions which exhibit an amplitude growing with the distance from metal - air boundary .
few such lowest modes were found numerically in ref .
@xcite but they were erroneously identified as surface modes .
we argue that these are the creeping modes . to find their dispersion relation ,
we assume that either @xmath172 or @xmath173 .
then eq .
reduces to @xmath174 where the upper sign corresponds to the case @xmath172 , and the lower sign to @xmath173 . noting that @xmath175 , where integer number @xmath176 runs from @xmath177 to @xmath178 , we obtain an infinite number of roots , @xmath179 in figs . [ fig : airycomplexmap ] , these roots are located on the ray @xmath120 and a few lowest roots are encircled by white circumferences . to the best of our knowledge ,
the dispersion relation was first derived in ref .
* eq . 141.1 ) .
let us now consider the wave of the second kind which is described by eq . and obeys the dispersion equation eq . .
as discussed in sec .
[ s3 ] , this wave , if exists , could be evanescent at @xmath149 according to eq .
provided that eq
. holds . in the range @xmath158 an appropriate asymptotes at for @xmath180 have the form ( * ? ? ?
9.7.5 and 9.7.5 , 9.7.6 , 9.7.13 , and 9.7.14 ) @xmath181 note the absence of subdominant terms proportional to @xmath182 . putting these asymptotes into eq .
yields the equation @xmath183 solving it by iterations for the case @xmath184 , we find @xmath185 where the second term in rhs again represents a radiative correction to the dispersion of planar sw .
. is valid if @xmath186 , and the last term is small in comparison to the first one .
it means that a surface wave of the second kind exists for any @xmath14 in the allowed range if @xmath130 .
this conclusion is supported by numerical simulations as reported in sec .
it is also found that sw of the second kind disappears for @xmath187 since @xmath188 and sw mode is transformed into a radially growing mode .
a critical value of @xmath96 can be in principle found for which @xmath189 , and it occurred to be a function of @xmath169 .
it is somewhat surprising that the last term in , which stands for the radiative correction to the dispersion relation of sw , diminishes @xmath190 .
it may reverse the sign of @xmath191 if @xmath192 and this conclusion is supported by numerical simulations : compare @xmath110 and @xmath193 in the caption to fig .
[ fig : airycomplexmapplus ] .
it is worth noting that a result similar to eq .
was obtained by j. wait @xcite although it is not clear which of the two equations and he used . in conclusion to this section , we note that serial roots are located near the ray @xmath131 and they are complex conjugate to similar roots found for the wave of the first kind .
reversing the sign of @xmath18 in eq .
yields @xmath194keeping the same notations and as for the convex case brings the last equation to the form @xmath195 instead of eq . .
its general solution reads @xmath196 using asymptotic formulas @xmath197 for large positive values of the argument , @xmath198 , leads to the conclusion that radially evanescent solution occurs if @xmath199 . the boundary condition then reads @xmath200 the asymptote for @xmath201 is valid in the entire range of @xmath202 except for a vicinity of the ray @xmath203 . taking also the asymptotic formula @xmath204 for the derivative of @xmath201 in the same range of @xmath202
we reduce eq . to the form @xmath205 for @xmath184 . solving it same method as we used for eq .
yields @xmath206 note that radiative ( last ) term in has a sign opposite to that in eq . .
hence , now this term enhance damping of sw in the direction of propagation .
numerical simulation shows that sw exists for any value of @xmath96 .
the effect of curvature in negligible as long as @xmath207 and is dominant if @xmath208 .
creeping modes are not found in the concave geometry .
main results of this paper can be recovered from a solution of the wave equation , which can be solved in terms on the hankel function of the first kind @xmath209 and the second kind @xmath210 respectively for the outgoing and incoming waves ; here @xmath211 and @xmath125 .
corresponding dispersion equations read @xmath212 and @xmath213 where the prime stands for the derivative with respect to the function argument . a connection with the content of our paper can de drawn from the asymptotic formulas for the hankel functions : @xmath214 , \end{gathered}\ ] ] where the upper sign corresponds to @xmath215 , the lower to @xmath216 , @xmath217 , @xmath218 and , hence , @xmath219 . in our ordering scheme , @xmath220 , @xmath221 , and @xmath222 . as to validity of eq .
, its derivation assumes that @xmath223 and @xmath224 . for @xmath225 ,
a different asymptotes should be used , namely @xmath226 .
\end{gathered}\ ] ] it can readily be derived from the debye asymptotes of the hankel functions ( see e.g. appendix of ref .
note that the summand @xmath227 in the exponent is small as compared to @xmath228 if @xmath225 , however it can be neglected only if @xmath229 , when it becomes smaller than @xmath230 .
its real part @xmath231 gives insignificant phase shift but its imaginary part @xmath232 is responsible for strong radial variation of the eigen waves . because of this summand , the hankel function @xmath215 of complex order with @xmath233 turns out to be fast growing function of its argument in the entire interval @xmath234 in particular , @xmath235 , \end{gathered}\ ] ] if @xmath236 .
all the creeping waves with the eigenvalues are radially rising modes in the interval . they are transformed into algebraically decaying modes with @xmath237 at @xmath238 . hence , a creeping mode number @xmath176 is peaked at @xmath239 hence , the creeping waves in the class of hankel functions @xmath215 of the first kind should be ignored when searching for sw localized at the metal - air interface . as to the wave @xmath216 of the second kind
, it exponentially decreases in the interval of radii provided that @xmath240 .
hence , it is suitable for a sw localized at the metal - air interface . as we can deduce from our study of eq . , its full - scale analog can have one and only one root with @xmath240 .
it is this root , which stands for sw . at large distances from the interface , @xmath241 with @xmath242 given by eq .
, this wave is transformed into freely propagating incoming cylindrical wave .
for this reason a solution containing @xmath216 was thrown away in refs .
we note however that this reason seemed not convincing in case if @xmath243 and the transformation occurs at exponentially low amplitude of the wave , @xmath244 .
in the majority of the literature on the subject of electromagnetic waves propagation in the vicinity of a conducting cylinder , the hankel functions of the second kind are discarded with the argumentation that they describe waves that asymptotically propagate toward the cylinder .
one of the main results of our paper is that we show that for a boundary condition corresponding to a lossy boundary , this is , in general , incorrect .
indeed , the hankel functions in this case have large complex indices , and their asymptotic behavior is more complex than just `` outgoing '' or `` incoming '' wave criterion .
we consider the distribution of the poynting vector in the vicinity of the cylinder , and make our conclusion on localization of a particular mode with account of the energy flow in the mode .
our approach uses a paraxial equation to describe a surface wave at a cylindrical metal - air interface in the low frequency limit with @xmath68 .
its solution is found in terms of the airy - fock functions of the first and second kind , avoiding in this way the complexities of dealing with the hankel functions of large argument and large order used in earlier studies .
we have shown that surface waves of two types can exist on a curved metal boundary which we called the outgoing and incoming surface waves . inspecting the radial component @xmath128 of the poynting vector , we reveal a qualitative difference between the two types of the surface waves . in the vicinity of the metal - air interface ,
the radial energy flux is directed inwards for the waves of both types as expected for a lossy medium ( ie , a realistic metal ) .
however @xmath128 reverses its sign at large distances from the interface in case of the outgoing surface wave .
we also emphasized that only incoming surface wave survives in the limit of planar metal - air boundary this fact , being always known , was not properly evaluated by those authors who discarded incoming solutions in the cylindrical geometry .
we have derived the dispersion relations for the waves of both types by computing the radiative corrections to the dispersion relation of a planar surface wave due to the curvature of the metal - air interface , see eqs . and .
we found that for any value of the surface impedance there is at most one surface mode of both types .
we believe that some of the modes previously reported as the surface waves are indeed the so called creeping waves .
the authors are grateful to b.a .
knyazev , i.v . andronov and fourth referee for useful references .
d. l. mills , v. m. agranovich , http://books.google.ru/books?id=wkjpv8c2ztqc[surface polaritons : electromagnetic waves at surfaces and interfaces ] , north - holland publ . , 1982 . http://books.google.ru/books?id=wkjpv8c2ztqc r. stanley , http://www.nature.com/nphoton/journal/v6/n7/full/nphoton.2012.161.html[plasmonics in the mid - infrared ] , nature photonics 6 ( 2012 ) 409411 .
http://dx.doi.org/10.1038/nphoton.2012.161 [ ] .
g. n. zhizhin , o. i. kapusta , m. a. moskaleva , v. g. nazin , v. a. yakovlev , http://iopscience.iop.org/0038-5670/18/11/a09/[spectroscopy of surface waves and the properties of the surface ] , sov .
uspekhi 18 ( 11 ) ( 1975 ) 927928 .
g. n. zhizhin , m. a. moskaleva , e. v. shomina , v. a. yakovlev , http://books.google.ru/books?id=wkjpv8c2ztqc[surface electromagnetic wave propagation on metal surfaces ] , north - holland publ . , 1982 ,
ch . 3 , pp .
h. raether , http://link.springer.com/chapter/10.1007/bfb0048319[surface plasmons on smooth and rough surfaces and on gratings ] , vol .
111 of springer tracts in modern physics , springer - verlag , new york , 1988 .
http://dx.doi.org/10.1007/bfb0048319 [ ] .
http://link.springer.com/chapter/10.1007/bfb0048319 a. v. zayats , i. i. smolyaninov , a. a. maradudin , http://www.sciencedirect.com/science/article/pii/s0370157304004600[nano-optics of surface plasmon polaritons ] , physics reports 408 ( 34 ) ( 2005 ) 131314 .
http://dx.doi.org/http://dx.doi.org/10.1016/j.physrep.2004.11.001 [ ] .
http://www.sciencedirect.com/science/article/pii/s0370157304004600 v. v. gerasimov , v. s. cherkassky , b. a. knyazev , i. a. kotelnikov , a. k. nikitin , g. n. zhizhin , study of surface plasmons travelling along straight and curved metal - dielectric interfaces : experiment and theory , in : 37th international conference on infrared , millimeter and terahertz waves , september 23 - 28 , 2012 .
report tue - a-4 - 4 , university of wollongong , australia , 2012 .
i. a. kotelnikov , v. v. gerasimov , b. a. knyazev , http://journals.aps.org/pra/abstract/10.1103/physreva.87.023828[diffraction of a surface wave on a conducting rectangular wedge ] , phys .
a 87 ( 2 ) ( 2013 ) 023828 .
g. bogomolov , g. zhizhin , a. nikitin , b. knyazev , http://www.sciencedirect.com/science/article/pii/s0168900208020755[geodesic elements to control terahertz surface plasmons ] , nucl .
instr . and meth . a 603 ( 1 ) ( 2009 ) 5255
. http://dx.doi.org/10.1016/j.nima.2008.12.158 [ ] .
m. gong , t .-
jeon , d. grischkowsky , http://www.opticsinfobase.org/oe/abstract.cfm?uri=oe-17-19-17088[thz surface wave collapse on coated metal surfaces ] , opt .
express 17 ( 19 ) ( 2009 ) 1708817101 .
http://dx.doi.org/10.1364/oe.17.017088 [ ] .
http://www.opticsinfobase.org/oe/abstract.cfm?uri=oe-17-19-17088 m. v. berry , http://stacks.iop.org/0305-4470/8/i=12/a=011[attenuation and focusing of electromagnetic surface waves rounding gentle bends ] , journal of physics a : mathematical and general 8 ( 12 ) ( 1975 ) 1952 .
u. schrter , a. dereux , http://link.aps.org/doi/10.1103/physrevb.64.125420[surface plasmon polaritons on metal cylinders with dielectric core ] , phys .
b 64 ( 2001 ) 125420 .
http://dx.doi.org/10.1103/physrevb.64.125420 [ ] .
k. hasegawa , j. u. nckel , m. deutsch , http://scitation.aip.org/content/aip/journal/apl/84/11/10.1063/1.1675940[surface plasmon polariton propagation around bends at a metaldielectric interface ] , appl .
84 ( 2004 ) 18351837 . http://dx.doi.org/10.1063/1.1675940 [ ] .
k. hasegawa , j. u. nckel , m. deutsch , http://link.aps.org/doi/10.1103/physreva.75.063816[curvature-induced radiation of surface plasmon polaritons propagating around bends ] , phys .
a 75 ( 2007 ) 063816 . http://dx.doi.org/10.1103/physreva.75.063816 [ ] . http://link.aps.org/doi/10.1103/physreva.75.063816 j .- w .
liaw , p .-
wu , http://www.opticsexpress.org/abstract.cfm?uri=oe-16-7-4945[dispersion relation of surface plasmon wave propagating along a curved metal - dielectric interface ] , opt .
express 16 ( 7 ) ( 2008 ) 49454951 .
http://www.opticsexpress.org/abstract.cfm?uri=oe-16-7-4945 m. guasoni , http://josab.osa.org/abstract.cfm?uri=josab-28-6-1396[analytical approximations of the dispersion relation of the plasmonic modes propagating around a curved dielectric - metal interface ] , j. opt .
b 28 ( 6 ) ( 2011 ) 13961403 .
http://dx.doi.org/10.1364/josab.28.001396 [ ] .
j. polanco , r. fitzgerald , a. maradudin , http://www.sciencedirect.com/science/article/pii/s0030401813011401[propagation of p - polarized surface plasmon polaritons circumferentially around a locally cylindrical surface ] , optics communications 316 ( 0 ) ( 2014 ) 120 126 . http://dx.doi.org/http://dx.doi.org/10.1016/j.optcom.2013.12.003 [ ] .
b. rulf , http://scitation.aip.org/content/aip/journal/jmp/8/9/10.1063/1.1705421[relation between creeping waves and lateral waves on a curved interface ] , j. math .
phys . 8 ( 9 ) ( 1967 ) 17851793 .
y. m. chen , http://scitationnew.aip.org/content/aip/journal/jmp/5/6/10.1063/1.1704183[diffraction by a smooth transparent object ] , j. math .
phys . 5 ( 6 ) ( 1964 ) 820832 .
http://dx.doi.org/http://dx.doi.org/10.1063/1.1704183 [ ] .
w. streifer , r. d. kodis , http://www.dtic.mil/cgi-bin/gettrdoc?ad=ad0418708[on the solution of a transcendental equation arising in the theory of scattering by a dielectric cylinder .
part ii ] , tech .
rep . , dtic document ( 1962 ) .
http://www.dtic.mil/cgi-bin/gettrdoc?ad=ad0418708 g. v. stupakov , i. a. kotelnikov , http://link.aps.org/doi/10.1103/physrevstab.6.034401[shielding and synchrotron radiation in toroidal waveguide ] , phys . rev .
beams 6 ( 2003 ) 034401 . http://dx.doi.org/10.1103/physrevstab.6.034401 [ ] .
http://link.aps.org/doi/10.1103/physrevstab.6.034401 g. v. stupakov , i. a. kotelnikov , http://link.aps.org/doi/10.1103/physrevstab.12.104401[calculation of coherent synchrotron radiation impedance using the mode expansion method ] , phys . rev .
beams 12 ( 2009 ) 104401 . http://dx.doi.org/10.1103/physrevstab.12.104401 [ ] .
http://link.aps.org/doi/10.1103/physrevstab.12.104401 r. paknys ,
n. wang , http://ieeexplore.ieee.org/xpl/articledetails.jsp?arnumber=1143865[creeping wave propagation constants and modal impedance for a dielectric coated cylinder ] , antennas and prop .
, ieee trans . 34 ( 5 ) ( 1986 ) 674680 . http://dx.doi.org/10.1109/tap.1986.1143865 [ ] .
i. andronov , d. bouche , http://www.tandfonline.com/doi/abs/10.1163/156939395x00406[the creeping and whispering gallery waves on the surface of a transparent body ] , j. of electr .
waves and applications 9 ( 4 ) ( 1995 ) 503520 .
http://dx.doi.org/10.1163/156939395x00406 [ ] .
r. paknys , d. r. jackson , http://ieeexplore.ieee.org/xpl/articledetails.jsp?arnumber=1406218[the relation between creeping waves , leaky waves , and surface waves ] , antennas and prop .
, ieee trans . 53 ( 3 ) ( 2005 ) 898907 .
http://ieeexplore.ieee.org/xpl/articledetails.jsp?arnumber=1406218 i. v. andronov , http://link.springer.com/article/10.1134/s1063771008020012[the influence of the impedance on creeping wave propagation ] , acoust .
54 ( 2 ) ( 2008 ) 139145 .
http://dx.doi.org/10.1134/s1063771008020012 [ ] .
j. polanco , r. fitzgerald , a. maradudin , http://www.sciencedirect.com/science/article/pii/s0375960112003362[propagation of s - polarized surface polaritons circumferentially around a locally cylindrical surface ] , physics letters a 376 ( 18 ) ( 2012 ) 1573 1575 .
m. a. leontovich , on a method of solving the problem of propagation of electromagnetic waves along the earth s surface , proceedings of the academy of sciences of ussr , physics 8 ( 1 ) ( 1944 ) 16 , ( in russian ) .
m. a. leontovich , approximate boundary conditions for the electromagnetic field on the surface of a good conductor , vol .
part ii , printing house of the ussr academy of sciences , moscow , 1948 , pp .
512 .
b. a. knyazev , g. n. kulipanov , n. a. vinokurov , http://iopscience.iop.org/0957-0233/21/5/054017[novosibirsk terahertz free electron laser : instrumentation development and experimental achievements ] , meas .
sci . and techn .
21 ( 2010 ) 054017 .
http://dx.doi.org/10.1088/0957-0233/21/5/054017 [ ] .
http://iopscience.iop.org/0957-0233/21/5/054017 v.
v. gerasimov , b. a. knyazev , i. a. kotelnikov , a. k. nikitin , v. s. cherkassky , g. n. kulipanov , g. n. zhizhin , http://www.opticsinfobase.org/josab/abstract.cfm?uri=josab-30-8-2182[surface plasmon polaritons launched using a terahertz free - electron laser : propagation along a gold zns air interface and decoupling to free waves at the surface edge ] , j. opt .
am . b 30 ( 8) ( 2013 ) 21822190 .
http://dx.doi.org/10.1364/josab.30.002182 [ ] .
http://www.opticsinfobase.org/josab/abstract.cfm?uri=josab-30-8-2182 f. w. j. olver , d. w. lozier , r. f. boisvert , c. w. clark , http://www.google.ru/books?id=3i15ph1qf38c[nist handbook of mathematical functions ] , cambridge university press , 2010 .
http://www.google.ru/books?id=3i15ph1qf38c j. r. wait , https://ieeexplore.ieee.org/xpl/articledetails.jsp?arnumber=1144862[on the excitation of electromagnetic surface waves on a curved surface ] , antennas and prop .
, ieee trans . 8 ( 4 ) ( 1960 ) 445448 .
n. logan , k. yee , http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=1137826[a simple expression for propagation constant associated with a reactive convex surface ] , antennas and prop .
, ieee trans . 10 ( 1 ) ( 1962 ) 103103 . | we study propagation of electromagnetic surface waves on a metal - air interface in the case when the wave frequency is below the plasma frequency .
we derive a reduced wave equation for a metal cylinder with a given radius of curvature . using the leontovich boundary condition we find solutions to this equation which we classify as outgoing and incoming surface waves .
we derive the dispersion relations of the surface waves of both types and argue that the earlier studies overlooked the waves of the second type although they are the only type which can propagate on a planar metal - air boundary . surface wave , surface plasmons , paraxial approximation , leontovich boundary conditions |
in this paper we consider some classes of algebras of bounded linear operators on banach spaces over non - archimedean fields .
firstly , in terms of a banach algebra generated by a single operator , we define its normality property . in section 2 , we prove the strong normality ( the definitions are given below ) of the orthoprojections ( in the non - archimedean sense ) and multiplication operators on function spaces .
we also consider commutative algebras of normal operators .
for other approaches in this direction see @xcite .
secondly ( section 3 ) we propose a possible way based on the notion of a baer ring @xcite ) to develop a counterpart of the notion of a von neumann algebra .
the idea is based on a reduction procedure .
given a non - archimedean operator algebra @xmath0 , this procedure produces a ring @xmath1 of operators on a vector space over the residue field ( see @xcite for various applications of this idea ) . for algebras @xmath0
whose reductions belong to the class of baer rings , this leads to a kind of classification .
short of any general theory of that kind , we consider in detail a non - archimedean version of the crossed product construction , one of the main methods of constructing von neumann algebras in the classical case ( see the original paper @xcite by murray and von neumann ; for a modern exposition see @xcite ) .
this results ( section 4 ) in a class of non - trivial non - archimedean factors ( algebras with a trivial center ) or algebras close to factors corresponding , through the reduction procedure , to type i baer rings .
in this section we recall the notions from @xcite and provide some corrections and additional material .
we will not explain the basic notions of non - archimedean analysis ; see @xcite .
_ spectral theorem .
_ let @xmath2 be a bounded linear operator on a banach space @xmath3 over a complete non - archimedean valued field @xmath4 with a nontrivial valuation ; @xmath5 will denote the absolute value in @xmath4 , @xmath6 is the ring of integers in @xmath4 .
we denote by @xmath7 both the norm in @xmath3 and the operator norm @xmath8 ( see @xcite for some subtleties regarding the operator norm in the non - archimedean case ) .
below we assume that @xmath4 is algebraically closed .
we will denote by @xmath9 the residue field of @xmath4 .
denote by @xmath10 the commutative banach algebra generated by @xmath2 and the unit operator @xmath11 .
@xmath10 is the closure of the algebra @xmath12 $ ] of polynomials in @xmath2 , with respect to the norm of operators ; thus @xmath10 is a banach subalgebra of the algebra @xmath13 of all bounded linear operators .
elements @xmath14 are identified with the operators @xmath15 .
the spectrum @xmath16 of the algebra @xmath10 is defined ( see @xcite ) as the set of all bounded multiplicative seminorms @xmath5 on @xmath10 .
@xmath16 is provided with the weakest topology , with respect to which all functions @xmath17 , @xmath18 , are continuous . in this topology , it is a nonempty hausdorff compact topological space .
if the algebra @xmath10 is uniform , that is @xmath19 for any @xmath20 , and all the characters take their values in @xmath4 , then @xcite the space @xmath16 is totally disconnected , and @xmath10 is isomorphic to the algebra @xmath21 of continuous functions on @xmath16 with values in @xmath4 . in this case
the above isomorphism transforms the characteristic functions @xmath22 of nonempty
open - closed subsets @xmath23 into idempotent operators @xmath24
, @xmath25 . these operators form a finitely additive norm - bounded projection - valued measure on the boolean algebra of open - closed sets , with the non - archimedean orthogonality property @xmath26 an operator with the above properties is called _
normal_. it is called _ strongly normal _ , if its spectrum @xmath27 is a nonempty totally disconnected compact subset of @xmath4 , and @xmath28 . for a strongly normal operator @xmath2 , we have the spectral decomposition @xmath29 for any @xmath30 we can define the operator @xmath31 the operator @xmath32 is strongly normal @xcite . in order to check the normality of an operator @xmath2
, one has to establish the equality @xmath33 ^ 2\|=\|q(a)\|^2\ ] ] for any polynomial @xmath34 over @xmath4 .
unfortunately , a simple sufficient condition for ( 2.1 ) proposed in @xcite has been found to be wrong .
a counter - example was constructed by mihara @xcite who noticed that the matrix @xmath35 over @xmath36 or @xmath37 satisfies the reduction condition from @xcite , though ( 2.1 ) is violated for @xmath38 .
in fact , the reduction condition from @xcite implies only the equality @xmath39 which is weaker than normality .
only this property is actually proved in @xcite for the operators given there as examples , except the operators related to the non - archimedean model of the canonical commutation relations of quantum mechanics , whose normality is either immediate from the definition , or follows from a result given below ( section 2.3 ) . .
a projection on a banach space @xmath40 is such a linear bounded operator @xmath41 that @xmath41 is idempotent : @xmath42 .
it is obvious that either @xmath43 , or @xmath44
. kernels of @xmath45 and @xmath46 complement each other having a trivial intersection ; if they are orthogonal ( in the non - archimedean sense ) , then @xmath41 is called an _ orthoprojection_. in this and only this case , @xmath47 . for an orthoprojection @xmath41 different from 0 and @xmath11 , @xmath48 ( @xcite , page 63 ) .
a projection @xmath41 is an orthoprojection , if and only if it is strongly normal .
note that the spectrum of an orthoprojection is contained in the two - point set @xmath49 .
indeed , consider the equation @xmath50 let us write @xmath51 where @xmath52 , @xmath53 , @xmath54 we can seek a solution of ( 2.2 ) in the form @xmath55 , @xmath56 , @xmath57 . then ( 2.2 ) means that @xmath58 , and if @xmath59 , we can write a solution as follows : @xmath60 by ( 2.3 ) , @xmath61 is surjective .
one verifies directly that @xmath61 is injective .
suppose that @xmath41 is an orthoprojection different from 0 or @xmath11 . in this case @xmath62 .
the banach algebra @xmath63 generated by @xmath41 and @xmath11 consists of operators @xmath64 , @xmath65 , or , equivalently , of operators @xmath66 , @xmath67 , for which @xmath68 let us take such vectors @xmath56 , @xmath57 , that @xmath69 , and set @xmath55 . due to the orthogonality , @xmath70 . on the other hand , @xmath71x = ag+bh$ ] , so that @xmath72x\|=\max ( |a|,|b|)\cdot \|x\|.\ ] ] together with ( 2.4 ) , this implies the equality @xmath73 it follows from ( 2.5 ) that @xmath63 is isomorphic to @xmath74 .
hence @xmath41 is strongly normal .
conversely , if @xmath41 is a strongly normal projection , then under the isomorphism of @xmath63 and @xmath75 , the operator @xmath41 corresponds to such a continuous function @xmath76 that @xmath77 , so that @xmath78 where @xmath79 .
if @xmath45 , then @xmath80 is nonempty , so that @xmath81 , and @xmath41 is an orthoprojection .
@xmath82 .
let us consider the banach space @xmath83 where @xmath84 is a compact totally disconnected hausdorff topological space .
it is known that @xmath40 possesses an orthonormal basis ( @xcite , corollary 5.25 , or @xcite , theorem 2.5.22 ) .
let @xmath2 be an operator of multiplication on @xmath40 by a function @xmath85 .
then @xmath2 is strongly normal . .
obviously , the spectrum @xmath27 coincides with the set @xmath86 of values of the function @xmath87 .
the set @xmath88 is a compact subset of the zero - dimensional topological field @xmath4 .
therefore ( @xcite , 6.2.10 and 6.2.11 ) @xmath88 is zero - dimensional , hence totally disconnected .
the banach algebra @xmath10 can be identified with the completion of the set of functions @xmath89 , @xmath90 , where @xmath91 $ ] , with respect to the norm @xmath92 by kaplansky s theorem ( theorem 43.3 in @xcite ) , the algebra @xmath10 coincides with the set of operators of multiplication by functions @xmath93 , @xmath94 , thus it is isomorphic to @xmath95 .
* 3.1 . * _ commutative algebras_. classically , if a @xmath97-algebra @xmath0 of bounded operators on a hilbert space is commutative , then @xmath98 for any @xmath99 , so that all the operators from @xmath0 are normal .
therefore in our situation it is reasonable to consider a commutative algebra @xmath0 of normal operators on a banach space @xmath40 over the field @xmath4 .
we assume that @xmath0 is complete with respect to the norm of operators and contains the unit operator @xmath11 . under the above assumptions , the algebra @xmath0 is isomorphic to the algebra @xmath100 of @xmath4-valued continuous functions on a compact totally disconnected hausdorff topological space @xmath84 .
under this isomorphism , characteristic functions of open - closed subsets @xmath101 correspond to orthoprojections @xmath102 forming an orthoprojection - valued finitely additive measure on the algebra of open - closed subsets of @xmath84 . for an operator @xmath103 corresponding to a function @xmath104
, there is an integral representation @xmath105 convergent with respect to the norm of operators . .
for any operator @xmath99 , its normality means that the algebra @xmath10 is uniform and all its characters take values in @xmath4 .
then the whole algebra @xmath0 possesses the same properties ; therefore @xcite @xmath0 is isomorphic to @xmath106 .
the proof of ( 3.1 ) is identical to that of the representation of a single normal operator @xcite .
@xmath96 . in order to introduce a possible non - archimedean counterpart for the class of von neumann algebras , we need the following reduction procedure .
we assume that the space @xmath40 is infinite - dimensional and possesses an orthonormal basis ( in the non - archimedean sense ; see @xcite ) .
then @xmath40 is isomorphic to the space @xmath107 of sequences @xmath108 , @xmath109 , @xmath110 , @xmath111 , by the filter of complements to finite subsets of an infinite set @xmath112 .
we may assume that @xmath113 .
let @xmath0 be an algebra of linear bounded operators on @xmath107 .
denote by @xmath114 the closed unit ball in @xmath0 the set of all operators from @xmath0 with norm @xmath115 .
@xmath114 is an algebra over the ring @xmath6 , just as its ideal @xmath116 consisting of operators of norm @xmath117 .
the _ reduced algebra _
@xmath118 can be considered as a @xmath9-algebra .
now we can look for a class of @xmath9-algebras , for which there is a ( purely algebraic ) theory parallel to the theory of von neumann algebras .
then the class of algebras @xmath0 corresponding to @xmath119 from that class will be the desired one .
an algebraic theory of the above kind is the theory of baer rings and algebras developed by kaplansky @xcite .
a unital ring @xmath88 is called the baer ring , if each left ( or , equivalently , each right ) annihilator in @xmath88 is generated by an idempotent element .
this property was proved by baer @xcite for the ring of all endomorphisms of a vector space of an arbitrary dimension .
kaplansky @xcite proved it for any @xmath120-algebra ( the class of @xmath120-algebras is wider than the class of von neumann algebras ) .
a baer ring @xmath88 is called abelian , if all its idempotents are central , and dedekind finite , if @xmath121 implies @xmath122
. an idempotent @xmath123 is called abelian ( finite ) , if the baer ring @xmath124 is abelian ( respectively , dedekind finite ) . if @xmath125 and @xmath126 are central idempotents , we write @xmath127 , if @xmath128 .
an idempotent @xmath129 is called faithful , if the smallest of the central idempotents @xmath126 satisfying @xmath130 is equal to 1 .
kaplansky @xcite introduced the following types of baer rings .
a baer ring @xmath88 is of type i , if it has a faithful abelian idempotent .
it is of type ii , if it has a faithful finite idempotent , but no nonzero abelian idempotents , and of type iii , if it has no nonzero finite idempotents .
these classes are subdivided further into finite and infinite ones .
a typical example of a type i baer ring is the ring of all linear transformations of a vector space of countable dimension . under some additional conditions , an arbitrary baer factor of type
i is of this form @xcite .
for other examples of baer rings see @xcite .
the main result @xcite of the theory of baer rings is the unique decomposition of every baer ring into a direct sum of rings of the above types .
let us call an operator algebra @xmath0 _ an algebra with the baer reduction _
, if the reduced algebra @xmath1 is a baer ring .
it is well known ( see @xcite , lemma 2.5.1/3 ) that a finite system of elements of norm 1 in a non - archimedean normed space is orthonormal if and only if their reductions are linearly independent .
therefore * the operator ring @xmath114 with the baer reduction is an orthogonal sum of rings with reductions of types i , ii , and iii*. the simplest example of algebras with the baer reduction of type i is the algebra of all bounded operators on a banach space of countable type ( that is , @xmath107 with a countable set @xmath112 ) . the very existence of other operator algebras ( moreover , factors ) with this property is far from obvious . below
we present a class of such algebras . as a whole
, the study of various classes of operator algebras with the baer reduction seems a huge problem comparable with the whole theory of von neumann algebras .
* 4.1 . * _ analysis on product spaces_. let @xmath131 be a totally disconnected compact hausdorff topological space , @xmath132 be an abelian infinite second - countable totally disconnected compact hausdorff topological group acting transitively on @xmath131 by homeomorphisms . the action will be denoted as @xmath133 , @xmath134 , @xmath135 ( the group operation is written multiplicatively ) . below we construct and study some operators on @xmath136 , the space of continuous functions on @xmath137 with values in @xmath37 . here
@xmath138 is a prime number , @xmath37 is the completion of an algebraic closure of the field @xmath36 of @xmath138-adic numbers
. we will denote @xmath139 the absolute value in @xmath37 .
some features of our approach follow @xcite though the actual meaning of our objects is different .
note also ( in order to avoid confusion referring to various sources ) that in the class of compact hausdorff spaces the properties of total disconnectedness and zero dimensionality are equivalent .
we begin with the following easy property of the space @xmath140 . as usual , we denote by @xmath141 the commutant of an operator algebra @xmath0 over a banach space @xmath40 , that is the set of such bounded operators @xmath142 on @xmath40 that @xmath143 for each @xmath99 .
below we use a similar notation for the commutant of an arbitrary set of operators . .
obviously , @xmath148 .
let @xmath149 .
then , for every @xmath145 , @xmath150 , that is @xmath151 for any @xmath152 . for @xmath153 , we get @xmath154 , so that @xmath155 , @xmath156 , @xmath134
. therefore @xmath157 from now on , we assume that * the group @xmath132 is @xmath138-compatible*. this notion is defined as follows ( see @xcite ) .
denote by @xmath158 the set of all such natural numbers @xmath159 , for which there exists a subgroup @xmath160 with the property that @xmath161 has an element of order @xmath159 .
the @xmath138-compatibility means that @xmath162 ( an additional condition formulated in @xcite , sect .
9.j , for more general fields is satisfied automatically for @xmath37 ) .
a typical example is @xmath163 where @xmath164 is a prime different from @xmath138 .
it is well known ( see @xcite ) that @xmath138-compatible groups possess nice properties resembling those appearing in classical harmonic analysis . in particular
, @xmath132 possesses a @xmath37-valued haar measure @xmath165 . in the banach space @xmath166
, there is an orthonormal basis @xmath167 consisting of @xmath37-valued characters , that is maps @xmath168 , such that @xmath169 .
it is convenient to index the characters not by natural numbers but by elements of the dual group @xmath170 consisting of all @xmath37-valued characters or , equivalently , of all continuous homomorphisms @xmath171 where @xmath172 is the set of all roots of 1 in @xmath37 of orders prime to @xmath138 ( @xcite , p. 360 ) . for groups of this type
, @xmath170 is isomorphic also to the pontryagin dual ( @xcite , p. 350 ) .
the second - countability property of @xmath132 implies its metrizability and the countability of @xmath170 ( @xcite , theorem 24.15 ) . below we will use the fact that @xmath170 ( with the discrete topology ) is torsional , that is every finite subset of @xmath132 lies in a finite subgroup ( see corollary 9.15 in @xcite ) . returning to characters
@xmath173 we stress that @xmath174 , @xmath175 , is identical with @xmath176 , so that @xmath177 , for any @xmath135 . in particular , @xmath178 .
thus , @xmath179 , @xmath180 .
note also @xcite that the characters are orthonormal not only in the non - archimedean sense , but also in the integral sense , with respect to the @xmath37-valued haar measure @xmath165 : @xmath181 where @xmath182 is the kronecker symbol .
dealing with a function @xmath183 we can write @xmath184 where @xmath185 ( for the integration theory with non - archimedean - valued measures see @xcite ) .
it follows from ( 4.3 ) that the functions @xmath186 are continuous .
below we use the notation @xmath187 where @xmath188 by the filter of complements to finite sets in @xmath170 ( in such cases we will write @xmath189 ) ; see , for example , proposition 1.6 in @xcite . here and below we denote the supremum norms in various spaces of continuous functions by the same symbol @xmath190 .
let @xmath191 be a linear bounded operator on @xmath136 ( letters with a bar will denote operators on @xmath136 , to distinguish them from operators on @xmath140 ) .
then for any @xmath145 , @xmath192 , @xmath193 where @xmath194 , @xmath195 , as @xmath196 , @xmath197 define an operator @xmath198 on @xmath140 by setting @xmath199 . by ( 4.4 ) and ( 4.5 ) , this operator is linear and bounded . for each @xmath159 ,
@xmath200 in the strong operator topology . by ( 4.5 ) , @xmath201 it follows from ( 4.2 ) and ( 4.4 ) that @xmath202 the change in the order of summation is justified as follows . by theorem 3.8 from @xcite ( which remains valid in the vector case and for the summation over @xmath170 ) , it is sufficient to prove that , for any @xmath203 , there exists a finite subset @xmath204 , such that @xmath205 whenever @xmath206 . indeed , we know that @xmath207 . by ( 4.6 ) , @xmath208 .
choose a finite set @xmath84 in such a way that @xmath209 , for @xmath210 .
on the other hand , for a fixed @xmath159 , @xmath211 is the @xmath212-th coefficient of the expansion of the function @xmath213 on @xmath132 in the basis @xmath214 .
therefore we find such a finite subset @xmath215 that @xmath216 if @xmath217 .
thus , @xmath218 for all @xmath192 , if @xmath217 , so that @xmath219 meanwhile , @xmath220 therefore we obtain the required inequality , if we set @xmath221 .
the equalities ( 4.7 ) show that the correspondences @xmath222 , @xmath223 , agree with the usual rules of linear algebra .
similarly , sums and products of these infinite matrices correspond to the sums and products of operators .
let @xmath224 be a fixed function tending to zero at infinity .
denote by @xmath225 the banach space of functions @xmath226 with finite norm @xmath227 we call a function @xmath228 , where @xmath229 and the sum is indexed by a finite set of elements @xmath135 , _ a trigonometric polynomial _ on @xmath170 . .
suppose that @xmath230 .
given @xmath203 , choose a finite subgroup @xmath231 in such a way that @xmath232 ( that is possible , since the group @xmath170 is torsional ) .
let us consider the restriction of @xmath76 to the subgroup @xmath233 .
since @xmath233 is finite , this restriction can be expanded into a finite linear combination of characters on @xmath233 with values in @xmath172 .
this is a consequence of the fact that @xmath170 and all its subgroups are @xmath138-compatible ( @xcite , p. 361 ) .
note also that for such groups there is an isomorphism between @xmath234-valued and @xmath172-valued characters ( @xcite , p. 360 ) .
in addition , every character of @xmath233 can be extended to a character on @xmath170 ( @xcite , p. 350 ) . since @xmath235 , each character on @xmath170 has the form @xmath236 where @xmath135 is a fixed element .
thus , we have the representation @xmath237 where @xmath229 , @xmath238 so that @xmath239 for @xmath240 .
note that the dual space to the banach space @xmath247 of bounded @xmath37-valued functions on @xmath170 is @xmath248 ( see theorem 5.5.5 in @xcite ) .
therefore the above reasoning proves also that the set of all trigonometric polynomials is weakly dense in @xmath247 .
denote by @xmath253 the set of all the operators @xmath254 and @xmath255 , and by @xmath256 the set of all the operators @xmath257 and @xmath258 .
the closed linear hull ( with respect to the strong operator topology @xcite ) of the set @xmath253 will be denoted @xmath259 .
similarly , @xmath260 is the strongly closed linear hull of @xmath256 .
it follows from ( 4.9 ) that the mapping @xmath261 is a spatial isomorphism of the algebras @xmath259 and @xmath260 .
these algebras will be interpreted as non - archimedean crossed product algebras , at least for the case of @xmath138-compatible groups .
if @xmath223 is an arbitrary bounded operator , then by ( 4.10 ) , @xmath268 also we have @xmath269 , @xmath270 .
thus , if an operator @xmath191 commutes with all @xmath255 , then by lemma 4.1 , each operator @xmath198 is the operator of multiplication by a continuous function @xmath271 on @xmath131 . if @xmath272 , then , in addition , @xmath273 , and we find from ( 4.11 ) that for any @xmath152 , @xmath274 which implies the identity @xmath275 let us find a representation of operators from @xmath256 .
we have , for any @xmath145 , @xmath276 so that @xmath277 next , @xmath278 writing @xmath279 we get @xmath280 so that @xmath281 where @xmath282 is the operator of multiplication by the function @xmath283 , an appropriate coefficient from ( 4.14 ) .
let us fix an element @xmath284 . denote by @xmath285 the set of such @xmath175 that the equality @xmath286 implies the equality @xmath287 .
the set @xmath285 is a subgroup in @xmath170 .
if the action is free , that is @xmath286 implies @xmath288 , then @xmath289 . for @xmath290 , define a function @xmath291 as follows .
since the action of @xmath132 on @xmath131 is transitive , for each @xmath134 there exists an element @xmath135 , such that @xmath292 .
set @xmath293 . if also @xmath294 , @xmath295 , then @xmath296 , and @xmath297 , provided @xmath290 .
therefore the function @xmath298 is well - defined .
note also that @xmath299 and @xmath300 . as we have proved , an operator @xmath272 is represented by the matrix of multiplication operators corresponding to the functions @xmath271 on @xmath131 , satisfying ( 4.12 ) or , equivalently , the identity @xmath301 set @xmath302 . if @xmath303 , then there exists such an element @xmath135 that @xmath304 and @xmath305
it follows from ( 4.16 ) that in this case @xmath306 , so that @xmath307 for all @xmath134 .
therefore @xmath308 where @xmath309 .
the function @xmath298 also enables us to give a more explicit representation of the coefficients @xmath310 appearing in ( 4.15 ) .
we find , using the orthogonality relation ( 4.1 ) , that @xmath311 for any @xmath312 , using the invariance of the haar measure , we get @xmath313 and setting @xmath302 we find that @xmath314 where @xmath315 , if @xmath316 .
it follows directly from the definitions that @xmath327 .
since @xmath328 is closed in @xmath13 with respect to the strong operator topology , the equality @xmath323 will be proved if we show that each element from @xmath328 can be approximated in strong operator topology by linear combinations of operators from @xmath256 . as we know , every operator @xmath191 from @xmath328 has a matrix representation @xmath329 where @xmath330 is the operator of multiplication by the function ( 4.17 ) , in which @xmath331 for all @xmath332 ( here and below the letter @xmath333 denotes various positive constants ) , @xmath334 , as @xmath196 .
let @xmath335 , and @xmath336 be an operator with the same matrix elements as @xmath191 on a single `` diagonal '' @xmath337 and zeroes on the rest of matrix entries .
let us prove that @xmath338 in the strong operator topology , as @xmath339 .
let @xmath340 , @xmath341 . then @xmath342 if @xmath318 , and @xmath343 , if @xmath344 .
it is sufficient to consider the case where @xmath318 . for any @xmath203
, there exists such a finite set @xmath231 that @xmath345 , as @xmath346 . if @xmath347 , then @xmath348 and @xmath349 .
if @xmath350 , then @xmath351 , as @xmath212 is outside a finite set @xmath352 ( because @xmath353 belongs to the finite set @xmath233 ) .
the ultrametric property shows that the operator @xmath191 can be approximated in the strong operator topology by finite sums of operators @xmath336 .
it remains to prove that each @xmath336 can be approximated in the same topology by linear combinations of operators from @xmath256 .
thus , from now on , @xmath164 will be fixed .
set @xmath368 identifying an operator of multiplication by a certain function with that function and using ( 4.13 ) , ( 4.15 ) , and ( 4.19 ) , we get for the `` matrix elements '' the expression @xmath369 \eta_l(x),\ ] ] whence @xmath370\varphi_{j\cdot l^{-1}}^{(i)}(x)\eta_l(x)\right\rangle_{j\in { \widehat{g}}},\quad i=1,\ldots , k,\ ] ] and it follows from ( 4.20 ) that @xmath360 .
this proves ( i ) .
suppose that an operator @xmath375 from @xmath328 commutes with @xmath256 , that is with operators @xmath258 for all @xmath145 , and with operators @xmath257 for all @xmath251 .
by ( 4.13 ) , @xmath376 @xmath377 therefore the commutation of @xmath191 and @xmath257 for all @xmath251 means that @xmath378 if @xmath379 . on the other hand
, it follows from ( 4.12 ) and the transitivity of the action of @xmath132 on @xmath131 that @xmath380 , for each @xmath175 .
note that in the general situation of a non - free action of @xmath132 , @xmath259 and @xmath260 are `` almost '' factors their central elements have the matrix representation @xmath385 where @xmath386 are constants whose values coincide for @xmath387 .
idempotents in @xmath260 have the form @xmath375 where @xmath388 @xmath389 are elements from @xmath37 , such that @xmath390 the above idempotents are orthoprojections if and only if @xmath391 , for all @xmath392 .
the mapping @xmath397 , @xmath398 , is an isometric imbedding .
its image @xmath399 consists of those functions @xmath400 , for which @xmath401
if @xmath402 ( given such a function @xmath400 , one can restore @xmath76 setting @xmath403 where @xmath292 ) .
let @xmath404 . then in the expansion @xmath405 only the coefficients @xmath406 , @xmath290 , can be different from 0 . indeed , if @xmath407 , then there exists @xmath295 , such than @xmath408 , @xmath409
. then @xmath410 , @xmath411 , @xmath135 . for this @xmath176 , due to the invariance of the haar measure , @xmath412^{-1}\int\limits_g\varphi ( a)g_i(a^{-1})\,\mu ( da)=[g_i(b)]^{-1}c_i,\end{gathered}\ ] ] so that @xmath413 . .
let us use the orthonormal basis in @xmath136 given in lemma 4.3 .
the basis elements @xmath416 are indexed by pairs @xmath417 , @xmath290 , @xmath395 . if @xmath418 , then by ( 4.17 ) , @xmath419 where @xmath420 . writing @xmath421 for @xmath422 we can rewrite ( 4.21 ) as the matrix representation @xmath423 where @xmath424 the operators @xmath191 from the unit ball @xmath114 are those with @xmath425 .
it follows from ( 4.22 ) that the matrix representation of a reduced operator from @xmath1 has matrix elements @xmath426 where @xmath427 , @xmath428 if @xmath422 .
it is easy to check that a product of two such matrices corresponds to the product of the matrices given by the first factors on the right in ( 4.23 ) .
hence , the algebra @xmath1 is isomorphic to the algebra @xmath429 of column - finite matrices @xmath430 over @xmath414 , such that @xmath428 if @xmath422 .
such matrices can be interpreted as operators @xmath431 on the @xmath414-vector space @xmath432 of sequences @xmath433 containing only a finite number of nonzero elements , with the following additional property .
let @xmath434 consist of such sequences that @xmath435 for @xmath407 , while @xmath436 is the set of such sequences that @xmath435 for @xmath290 .
then the subspaces @xmath437 and @xmath436 are invariant for the operators @xmath431 .
conversely , if @xmath442 , then @xmath443 for @xmath444 , whenever @xmath445 .
taking @xmath446 , where @xmath387 is a fixed element , we find that @xmath428 for any @xmath387 .
the condition @xmath444 is equivalent to the property that @xmath422 for any @xmath387 .
similarly , we check that @xmath428 , if @xmath447 , @xmath422 .
obviously , @xmath448 , so that @xmath429 is isomorphic to the direct sum of the algebras of all linear operators on @xmath437 and @xmath436 respectively .. each of the latter algebras is a type i baer ring , which implies @xcite the same property of @xmath449 999 y. amice , interpolation @xmath138-adique , _ bull .
* 92 * ( 1964 ) , 117180 .
r. baer , _ linear algebra and projective geometry _ , academic press , new york , 1952 .
v. g. berkovich , _ spectral theory and analytic geometry over non - archimedean fields _ , ams , providence , 1990 .
a. borel , _ linear algebraic groups _ , w. a. benjamin , new york , 1969 . s. bosch , u. gntzer , and r. remmert , _ non - archimedean analysis _ , springer , berlin , 1984 .
o. bratteli and d. w. robinson , _ operator algebras and quantum statistical mechanics _ , vol . 1 , springer , new york , 1979 .
r. engelking , _ general topology _ ,
heldermann , berlin , 1989 .
e. hewitt and k. a. ross , _ abstract harmonic analysis _
, vol . 1 , springer , berlin , 1963 . s. s. holland , jr . , remarks on type i baer and baer-@xmath97 rings , _ j. algebra _ * 27 * ( 1973 ) , 516522 . s. s. holland , jr . ,
orthomodularity in infinite dimensions ; a theorem of m. solr , _ bull .
* 32 * ( 1995 ) , 205234 .
r. v. kadison and j. r. ringrose , _ fundamentals of the theory of operator algebras _ , vol .
1 , academic press , new york , 1983 .
i. kaplansky , projections in banach algebras , _ ann .
_ * 53 * ( 1951 ) , 235249 .
i. kaplansky , _ rings of operators _ , w. a. benjamin , new york , 1968 . s. katok , _ p - adic analysis compared with real _ , ams , providence , 2007 .
a. n. kochubei , _ analysis in positive characteristic _ , cambridge university press , cambridge , 2009 .
a. n. kochubei , non - archimedean normal operators , _ j. math . phys .
_ * 51 * ( 2010 ) , article 023526 .
s. ludkovsky and b. diarra , spectral integration and spectral theory for non - archimedean banach spaces , _ int . j. math . and math .
* 31 * ( 2002 ) , 421442 .
t. mihara , personal communication , june 29 , 2012 .
f. j. murray and j. von neumann , on rings of operators , _ ann .
* 37 * ( 1936 ) , 116229 .
l. narici , e. beckenstein , and g. bachman , _ functional analysis and valuation theory _ , marcel dekker , new york , 1971 . c. perez - garcia and w. h. schikhof ,
_ locally convex spaces over non - archimedean valued fields _ , cambridge university press , cambridge , 2010 .
a. m. robert , _ a course in @xmath138-adic analysis _ ,
springer , new york , 2000 .
a. van rooij , _ non - archimedean functional analysis _ ,
marcel dekker , new york , 1978 . w. h. schikhof , non - archimedean representations of compact groups , _ compositio math .
_ * 23 * ( 1971 ) , 215232 .
w. h. schikhof , _ ultrametric calculus _ , cambridge university press , cambridge , 1984 .
m. takesaki , _ theory of operator algebras _ , i - iii , springer , berlin , 2002 - 2003 .
k. g. wolfson , baer subrings of the ring of linear transformations , _ math .
z. _ * 75 * ( 1961 ) , 328332 .
k. g. wolfson , on type i baer factors , _ comm . algebra _
* 27 * ( 1999 ) , 26532664 . | we study some classes of algebras of operators on non - archimedean banach spaces . in particular , we propose a non - archimedean version of the crossed product construction .
[ section ] [ section ] [ section ] |
coronal mass ejections ( cmes ) are large - scale structures carrying plasma and magnetic field from the sun . cmes are associated with flares , filament eruptions , shocks , radio bursts , solar energetic particle ( sep ) events , and have been found to be the primary cause of major geomagnetic disturbances and changes in the solar wind flow ( see , _ e.g. _ , schwenn _ et al .
_ , @xcite ; vilmer _ et al .
_ , @xcite ; mann _ et al .
_ , @xcite ; gopalswamy _ et al .
_ , @xcite ; leblanc _ et al .
_ , @xcite ; gosling , @xcite ) .
radio type iii bursts are fast frequency - drifting features that can be used as tracers of energetic electrons as they are caused by subrelativistic ( @xmath02.5 to @xmath080 kev ) electron streams travelling outward from the sun at speeds of about 0.1c to 0.5c ( see , _ e.g. _ , wild , @xcite ; dulk , @xcite ) . near - relativistic electrons ( 30315 kev ) can be observed _ in situ _ during sep events , in association with type iii bursts that continue into the ip space ( simnett @xcite ) .
radio type ii bursts , on the other hand , are thought to be caused by mhd shock waves propagating through the corona and interplanetary medium ( _ e.g. _ , nelson and melrose , @xcite ) .
some studies suggest that coronal metric type ii bursts and interplanetary ( ip ) decameter
hectometer ( dh ) type ii bursts do not have the same driver ( _ e.g. _ , cane and erickson , @xcite ) , and that the latter are exclusively created by the cme bow shocks ( _ e.g. _ , cliver , webb , and howard , @xcite ) . seps are believed to be accelerated in the lift - off phase of cmes , which could include either acceleration in flaring processes ( _ e.g. _ , de jager , @xcite ) , acceleration in coronal / interplanetary shocks ( _ e.g. _ , reames , @xcite and references therein ) , or in both ( _ e.g. _ , klein and trottet , @xcite ; klein _ et al .
_ , @xcite ; kocharov and torsti , @xcite ) .
recently , the role of cme interaction in sep production has also been discussed ( _ e.g. _ , gopalswamy _ et al .
_ , @xcite ; kahler , @xcite ; wang _ et al .
_ , @xcite , and references therein ) . one of the difficulties in determining the relevant processes lies in the uncertainties of timing .
seps are typically released no earlier than the maxima of the flare impulsive phases ( kahler , @xcite ) , yet at the time of sep release the height of the associated cme can vary from one to over ten solar radii ( kahler , @xcite ; krucker and lin , @xcite ; huttunen - heikinmaa , valtonen , and laitinen , @xcite ) .
it has also been observed that different particle species can be released non - simultaneously ( huttunen - heikinmaa , valtonen , and laitinen , @xcite ; krucker and lin , @xcite ; mewaldt _ et al .
_ , @xcite ) .
the _ in - situ _ observations of seps are strongly influenced by the magnetic connection between the spacecraft and the flare site ( _ e.g. _ , cane , reames , and von rosenvinge , @xcite ) .
for example , the nominally well - connected area for the spacecraft in the l1-point is around 60@xmath1 west on the solar hemisphere
. found that at larger angular distances between the observer s magnetic footpoint and the flare location , the onsets of energetic particles in relation to the microwave maximum are more delayed , and the proton onsets are relatively more delayed than the electron onsets . assuming that seps originated from the flaring processes , concluded that the propagation of @xmath020 mev protons from the flare site to the connecting magnetic field line is systematically slower than of @xmath00.5 mev electrons , and the electron - to - proton ratio tends to increase with angular distance from the flare site
of course , the modern view of particle acceleration in the solar corona does not rely on flares as the only source of energetic particles during sep events .
thus , instead of reflecting the azimuthal particle propagation conditions in the corona , the findings of kallenrode can be re - interpreted as reflecting the azimuthal and/or temporal evolution of the source , usually a coronal shock wave driven / launched by a cme ; a delay between the plasma eruption and the observed onset of the sep event reflects the establishment of the magnetic connection between the source and the observer and/or the efficiency of the acceleration and escape of particles in different parts ( bow / flanks ) of the global coronal shock wave .
consistent with the current views and the conclusions of are the results of , who found that in some events the injection time of the electrons is consistent with the timing of the type iii radio emission , while in other cases the electrons were apparently released up to half an hour later than the start of the type iii burst .
the non - delayed electron events were concentrated in the well - connected area , while the delayed events were more spread out .
similar delays were also observed between protons and helium nuclei by .
the near - simultaneous proton and helium events originated from the well - connected area , and the events with delayed helium nuclei were more spread out . in this paper
we study the possible sep accelerators on 9 november 2002 . in section 2
we analyse the multiwavelength observations of this event : in 2.1 we describe how sep onset times were estimated , in 2.2 we analyse the flaring activity and cmes near the sep onset times , and in 2.3 we identify accelerated particles via radio emission . in section 3
we calculate heights for the radio features and compare them with other structures .
results are presented in section 4 and they are further discussed in section 5 .
the energetic particle data for protons were recorded by soho / erne @xcite , and for electrons by _ wind_/3dp @xcite and soho / ephin @xcite .
white - light coronograph images were provided by soho / lasco @xcite , euv images by soho / eit @xcite , hard x - ray data by rhessi @xcite , and solar magnetograms by soho / mdi @xcite .
the radio imaging at meter wavelengths was done by the nanay radioheliograph in france ( nrh ; kerdraon and delouis , @xcite ) .
radio spectral data at decimetric
metric wavelengths were provided by artemis - iv in greece @xcite , phoenix-2 in switzerland @xcite , and nanay decameter array in france ( dam ; lecacheux , @xcite ) . at decameter
hectometer wavelengths the radio spectral data were obtained from the waves experiment @xcite onboard the _ wind _ satellite .
an observable signature of sep events is a velocity dispersion . assuming that particles with different energies are released simultaneously at or close to the sun , the onset of the event at a distance from the sun
should be observed earlier at higher energies than at lower ones .
velocity dispersion analysis makes it possible to infer the release time of particles at the sun and the apparent path length travelled .
the velocity dispersion analysis has been utilized by several authors ( _ e.g. _ , debrunner , flueckiger , and lockwood , @xcite ; torsti _ et al .
_ , @xcite ; krucker _ et al .
_ , @xcite ; mewaldt _ et al .
_ , @xcite ; tylka _ et al .
_ , @xcite ; dalla _ et al .
_ , @xcite ; hilchenbach _ et al .
_ , @xcite ) .
recently , the validity of the velocity dispersion analysis has been examined by simulation studies ( lintunen and vainio , @xcite ; saiz _ et al .
_ , @xcite ) . in this study
, we determine the _ in - situ _ onset times of sep events by adopting the method used by .
the release times at the sun are then first derived using the velocity dispersion analysis . however , since the analysis has several complex systematic errors , we finally estimate the temporal windows for the particle release at the sun using a relaxed condition for the path length . for more detailed discussions on deriving the release times at the sun
, we refer to and , and references therein .
the velocity dispersion equation is @xmath2 where @xmath3 is the observed onset time at energy e at the spacecraft , @xmath4 is the release time from the acceleration site , @xmath5 is the apparent path length travelled by the particles , and @xmath6 is the velocity of the particle of energy @xmath7 .
the parameters are determined by fitting a straight line to the points @xmath8 . to make the release time derived from equation ( 1 ) comparable to the electromagnetic observations near earth
, eight minutes have to be added ( @xmath9=@xmath4 + 8 min ) .
the velocity dispersion analysis of the soho / erne proton data in the 1480 mev energy range yielded a path length of [email protected] au .
this energy range is covered entirely by the high energy detector ( he d ) of the erne instrument . with this path length
the proton acceleration was estimated to have taken place at 14:18 ut @xmath10 12 minutes .
the proton count rates , at several energy channels shifted back in time to compensate for the velocity dispersion , are presented in figure [ fig:1 ] . for details.,scaledwidth=100.0% ] + using _
wind_/3dp ( 0.0200.225 mev ) and soho / ephin ( 0.253.00 mev ) _ in situ _ electron observations and assuming the same path length as for the protons , the electron acceleration is estimated to have started at 13:53 ut @xmath10 11 minutes ( figure [ fig:2 ] ) .
however , the path length of 1.03 au seems to fail in compensating the velocity dispersion of the electron onsets .
the velocity dispersion analysis performed only on the electron data yields a very high path length of [email protected] au , which we deem somewhat unreliable ( figure [ fig:3 ] ) . with this high path length
the electron release times would have been earlier , at 13:31 ut @xmath10 8 minutes , but most likely the high value of the path length is caused by an inaccurate determination of the onset times in the low - energy electron channels due to the long rise time and high background level of the fluxes .
more conservative estimates of the release time windows are evaluated using path lengths in the 1.01.4 au range . to evaluate the uncertainties in the release times , the standard deviation error limits ( 1@xmath11 ) for the average proton and electron release times
were calculated and used to determine the release time windows , as shown in figure [ fig:3 ] .
it is clear that protons are delayed with respect to electrons .
the delay is much larger if we accept the very high path lengths for the electrons obtained from the velocity dispersion fit to the electron data only .
based on figure [ fig:3 ] , we estimate that the protons in this event were released between 13:57 and 14:27 ut , and the electrons earlier , between 13:40 and 14:03 ut , assuming a low path length for both populations .
the electrons could have been released even earlier than that , around 13:2313:39 ut , if a much higher path length is accepted for the electron population . on 9 november 2002 , during the estimated sep release period , two cmes were observed .
the first ( cme1 in figure [ fig:4 ] ) was directed to the northwest .
it was first detected at 11:06 ut in lasco c2 , at a height of 2.55 r@xmath12 , travelling at a plane - of - the - sky speed of 530 km s@xmath13 ( determined from a linear fit to all height - time data points from the lasco cme catalog ) .
the second cme ( cme2 in figure [ fig:4 ] ) , was halo - type and directed to the southwest , and it was first detected at 13:31 ut at a height of 4.20 r@xmath12 .
the plane - of - the - sky speed from a linear fit to all data points was 1840 km s@xmath13 .
a second - order fit to the same data points indicates that the halo cme was accelerating , reaching a speed of 2000 km s@xmath13 near 30 r@xmath12 .
we estimate that the lateral expansion speed ( see schwenn _
et al . _ , @xcite ) was even higher than that , about 2300 km s@xmath13 before 14:00 ut , and about 1800 km s@xmath13 after 14:00 ut .
this speed is , however , difficult to measure , since the northwestern cme2 edge was irregular and varying in time .
the three consecutive lasco c2 difference images in figure [ fig:4 ] show how the two cmes evolved and how the halo cme expanded laterally . at 13:54 ut ( middle panel in figure [ fig:4 ] )
an arc - like brightening is visible between the two cmes . in the later lasco c3 images
the halo cme surrounds the disk in full 360 degrees . .
the later halo - type cme ( cme2 ) , was directed to the southwest and had a plane - of - the - sky speed of about 1840 km s@xmath13 .
the white dashed line ( indicated with an arrow ) outlines a brightening on the side of the halo cme at 13:54 ut .
, scaledwidth=100.0% ] the height time plot in figure [ fig:5 ] shows the estimated heights of the cme leading edges , combined with the goes soft x - ray flux curve during 10:3015:00 ut . the backward - extrapolated height times for cme2 give an estimated start time for the halo cme near 13:05 ut .
the fast halo cme2 was clearly associated with activity in noaa ar 0180 at s11w36 ( location on 9 november at 12:00 ut ) .
the active regions most probably associated with the earlier cme1 were ar 0177 at n18w30 ( location on 7
november at 06:30 ut ) , and ar 0175 at n15w81 ( location on 9 november at 15:30 ut ) .
these active regions are shown in an eit map on 7 november , in figure [ fig:6 ] .
the difference in latitude between the source regions of the two cmes was 2530@xmath1 and the difference in longitude 2040@xmath1 , depending on the origin of cme1 ( ar 0175 or ar 0177 ) .
since cme2 developed into a full halo and showed lateral expansion , it is possible that the two cmes made contact . however , as cmes are observed as projections on the plane of the sky , determining the real geometry is prone to uncertainties . to the west .
bottom : potential field lines from magnetic field extrapolations ( pfss ) using soho / mdi magnetogram on 9 november 2002 at 12:04 ut .
field lines plotted in magenta are open , with negative polarity .
, title="fig:",scaledwidth=75.0% ] to the west .
bottom : potential field lines from magnetic field extrapolations ( pfss ) using soho / mdi magnetogram on 9 november 2002 at 12:04 ut .
field lines plotted in magenta are open , with negative polarity .
, title="fig:",scaledwidth=55.0% ] in each frame.,scaledwidth=100.0% ] several goes c - class flares were recorded during the first cme passage , before 13:00 ut on 9 november , but for many of them the location was not identified . at 13:08 ut a goes m4.6 class flare started in ar 0180 , at s12w29 .
the flux maximum occurred at 13:23 ut and the flux decayed to the pre - flare level at about 15:00 ut .
the timing indicates clear association with the fast halo cme2 .
rhessi observed the flare in hard x - rays until 13:28 ut , when satellite night - time set in .
the highest energy band in which the flare was observed was 100300 kev . according to the rhessi flare list
the flare was located at @xmath14 , @xmath15 from the disk center .
after the satellite night - time ended , rhessi observed an x - ray brightening at 14:05 ut , slightly off from the earlier flare site , at @xmath16 , @xmath17 . the potential field ( _ i.e. _ , the current - free field ) extrapolations provide an approximation of the magnetic fields in a quiet sun .
however , the lack of electric current and free energy in the model means that it does not describe active regions very well .
the pfss solarsoft package @xcite uses the line - of - sight magnetograms provided by soho / mdi , and the potential field in figure [ fig:6 ] describes the situation on 9 november . on the east side of ar 0180
there are some open field lines ( plotted in magenta , indicating negative polarity ) , and many more in the regions between ar 0180 and ar 0177 .
it should be noted that the pfss maps describe the potential field prior to the event , and that ongoing events can strongly perturbe the magnetic structures . in euv ,
coronal loops were observed to open near the nortwest limb at 09:48 ut , at the same latitude where ar 0175 and ar 0177 were located .
an euv brightening appeared in this region at 12:24 ut , and in the following eit image bright material was ejected over the limb .
these transients were observed well before the halo cme2 appeared .
euv ribbons were formed over the ar 0180 ( s12w29 ) in the eit image at 13:13 ut , and they were clearly associated with the goes m4.6 flare .
post - flare loops appeared at 13:25 ut and an eit wave was observed in the same image . at
first it showed as a bright `` rim '' south of ar 0180 , and then as dark dimmed structures near the southwest limb after 13:36 ut ( figure [ fig:7 ] ) .
the dark structures persisted well after 14:12 ut .
a small euv brightening was observed in ar 0180 at 14:12 ut , possibly associated with the rhessi hard x - ray brightening at 14:05 ut . with radio observations we can trace accelerated particles and observe propagating shock fronts .
we therefore studied radio emission during 13:0015:00 ut at a large wavelength range ( 4 ghz100 khz ) .
the dynamic radio spectrum from artemis - iv at 100600 mhz shows narrow - band fluctuations ( nbf ) starting around 13:09 ut , one minute after the goes m4.6 flare start .
the fluctuations occured at metric wavelengths , in the 300100 mhz frequency range .
the fluctuating envelope drifted slowly towards the lower frequencies ( figure [ fig:8 ] ) .
the emission source locations could be imaged by nrh at 164 mhz , and figure [ fig:9 ] ( left panel ) shows how the emission centroids were located south of ar 0180 , along and over closed loop structures .
another drifting feature ( df ) was observed to start around 13:14 ut just below 500 mhz , with a frequency drift around 0.8 mhz s@xmath13 . the dashed lines in figure [ fig:8 ]
mark the frequencies that could be imaged with nrh , and figure [ fig:9 ] shows the source locations at the three different frequencies .
the source trajectory indicates that the burst driver was moving southward from the ar , with a slight path curvature towards the east .
we estimate from this image that the projected speed of the emission source was around 12001400 km s@xmath13
. shows some of these source locations , imaged by nrh ( times and frequencies are shown here with a white solid line and white dashed lines , respectively ) .
after 13:17:30 ut the features become unclear due to a group of metric type iii bursts superposed on the drifting emission.,scaledwidth=90.0% ] ) , through 327 mhz ( marked with + ) , to 236 mhz ( marked with @xmath18 ) .
the source positions indicate that the burst driver is moving southward from the ar , with a slight path curvature towards the east .
( the imaging periods were indicated in the dynamic spectrum in figure [ fig:8 ] . ) , scaledwidth=90.0% ] the_wind_/waves observations at 14 mhz20 khz ( figure [ fig:10 ] ) show an ip type ii burst lane at @xmath013:20 ut onwards .
there is some indication that the burst starts at a frequency higher than the rad2 upper limit at 14 mhz , as we see emission lanes also around 30 mhz ( ground - based observations ) .
during 13:3013:40 ut burst - like emission is observed also at 3565 mhz , and this enhanced emission looks to be separate from the type ii burst lane .
the type ii emission drifts down to 2 mhz by 13:48 ut , when it disappears under the stronger ip type iii burst groups .
the type ii burst lane appears occasionally also after 14:00 ut , and it is seen by waves rad1 at 500 khz near 15:20 ut . several metric dh type iii bursts were observed in association with the main impulsive phase of the m4.6 flare .
the first type iii burst group that appeared in the estimated time - frame for particle acceleration was observed at 13:28 ut , but the starting frequency is overshadowed by other burst structures .
the next group of type iii bursts started around 13:48 ut , with starting frequency near @xmath060 mhz ( figure [ fig:10 ] ; this frequency was also verified from the dam dynamic spectrum at 2070 mhz , not shown here ) .
another type iii burst group followed at 14:02 ut . within this group
burst lanes were observed to start at metric wavelengths near 250 mhz at @xmath014:05 ut ( frequency also verified from the phoenix-2 spectrum ) . ) . both of these features could have a continuation at frequencies lower than 70 mhz . at 13:2813:31 ut a type iii burst group
is observed at dh wavelengths , but the starting frequency is not clear . during 13:3013:40
ut enhanced burst emission is visible at 3565 mhz . around
13:48 ut a group of type iii bursts starts near 60 mhz , followed by another group around 14:05 ut around 250 mhz .
an ip type ii burst becomes visible around 13:23 ut at 14 mhz .
there is some indication that the burst could have started earlier at higher frequencies , but this emission is mixed with type iii bursts .
the type ii emission lane can easily be followed down to 2 mhz near 14:00 ut , when the more intense type iii bursts prevent seeing the fainter lane . estimated electron and proton release times for both path lengths
are also marked in the plot .
solid lines represent the release times using 1.01.4 au path lengths and dash - dotted line path lengths @xmath192.1 au , see section 2.1 for details . ] on closer inspection , the type iii bursts that started near 250 mhz look to be bidirectional .
the burst emission drifts both upward and downward in frequency ( figure [ fig:11 ] ) .
bursts with a negative frequency - drift continue to dh wavelengths , while the bursts with a positive drift can not be identified in the spectrum at frequencies higher than 350 mhz .
negative and positive frequency drifts are usually interpreted as electron beams travelling up and down in the corona , respectively .
some correlation can be found between these bursts and the count - rate peaks recorded by rhessi in hard x - rays ( figure [ fig:11 ] ) . since hard x - rays
are usually created by downward directed electron beams when they hit denser regions , the type iii bursts with a positive frequency drift could be tracing these beams .
the type iii bursts could be imaged at 327 , 236 , 164 , and 150 mhz by nrh , and the locations along one type iii burst are shown in figure [ fig:11 ] . the spatial separation between the rhessi x - ray burst and the type iii burst trajectory
indicate that either the sources were separate or the field lines were twisted . however , since the halo cme originated from this same active region , the lateral expansion of the cme could have affected the field lines .
we also note the vicinity of the open field lines on the east side of the active region ( shown in figures [ fig:6 ] and [ fig:11 ] ) .
all the observed type iii bursts continued to dh wavelengths , and the burst lanes look `` tilted '' at the time when they cross the type ii burst lane , see figure [ fig:12 ] .
for the 13:48 ut type iii burst group the tilt is not so obvious , but there is a brightening in the burst emission . for the 14:05 and 14:40 ut burst groups the type iii tilt is more pronounced .
it should be noted that in between 14:05 and 14:40 ut , _
i.e. _ , in the time frame for the estimated proton acceleration , no other burst activity was observed in the dm
dh spectra .
the radio emission during the flare
cme event showed several frequency - drifting structures .
as plasma emission is directly related to electron density , @xmath20=9000@xmath21 ( @xmath22 in @xmath23 and @xmath20 in hz ) , and electron density can be converted to atmospheric height with atmospheric density models , we calculate estimates for the radio source heights .
these heights and deduced speeds need to be considered critically , as they depend strongly on the selected density models .
selection of model should be based on the knowledge of coronal conditions in the propagation region , but the available information is often insufficient , see .
standard coronal density models such as and , and `` hybrid '' models such as , which is a mixture of the five - fold saito model for coronal densities and the model for ip densities , can be used to describe the undisturbed high corona at frequencies lower than @xmath0200 mhz ( corresponding electron density @xmath22 @xmath24 4@xmath2510@xmath26 @xmath23 ) . for the high - frequency emission at 500400 mhz ( corresponding electron density @xmath22 @xmath0 23@xmath2510@xmath27 @xmath23 ) , we need high - density models such as the ten - fold saito or a scale height method , to describe densities in active region loops or streamer regions . however , if the exiter of emission is propagating in the wake of earlier transients , in disturbed medium , the height estimates may not give reliable results .
the 300100 mhz frequency range of the burst envelope of the narrow - band fluctuations ( nbf ) suggests a relatively large height for the emission source already at 13:09 ut .
the estimated heliocentric distance at 100 mhz is 1.131.56 r@xmath12 , but the actual height value depends strongly on the density model used . from the radio spectrum in figure [ fig:10 ]
it is not clear if the fluctuating emission near 100 mhz is plasma emission at the fundamental frequency or at the second harmonic .
fundamental emission at 50 mhz would place the source height even higher in the corona .
the drifting emission feature ( df ) , observed at 400 mhz around 13:15 ut , had a height of 1.08 r@xmath12 according to the 10-fold saito density model . at 13:17 ut
the emission had drifted to 300 mhz , which corresponds to a height of 1.14 r@xmath12 using the same model .
the deduced burst driver speed is approximately 500 km s@xmath13 .
the speed values calculated directly from the atmospheric models assume that the source motion is directed radially against the density gradient , and the speed has to be multiplied by 1/cos(@xmath28 ) if these directions are separated by an angle @xmath28 .
usually 45@xmath1 and multiplication by 1.4 is considered as the realistic maximum correction needed .
the scale height method , using the observed @xmath00.8 mhz s@xmath13 frequency drift rate , yields a higher speed around 1420 km s@xmath13 .
we note that the burst sources of this feature , observed in the nrh images ( shown in figure [ fig:9 ] ) suggest a projected source velocity of 12001400 km s@xmath13 , which is close to the value obtained with the scale height method . the dh type ii burst , observed from 14 mhz at 13:23 ut to 2 mhz at 13:47 ut , moved from a heliocentric height of 2.78 r@xmath12 to 7.07 r@xmath12 ( vrnak , magdaleni , and zlopec hybrid density model ) .
the average speed along the burst lane was thus about 2070 km s@xmath13 . using the saito density model , instead of the hybrid , the heights are 2.07 and 5.18 r@xmath12 , with a speed of 1500 km s@xmath13 . in white light , the halo cme front was observed at a height of 5.89 r@xmath12 at 13:42 ut , which is closer to the heights calculated with the hybrid density model .
the observed start frequency of the second type iii burst group , around 60 mhz near 13:48 ut , suggests that electrons were accelerated near heliocentric distances 1.31.8 r@xmath12 ( with all the above - mentioned density models considered ) .
the third type iii group started at a much higher frequency , near 250 mhz at 14:05 ut , with bursts that showed both positive and negative frequency drifts .
this high frequency indicates a low heliocentric height for acceleration , @xmath291.2 r@xmath12 , with all the density models considered .
the sequence of the observed features is listed in table [ tbl1 ] .
the narrow - band fluctuations inside the drifting envelope started at 13:09 ut , which is during the flare impulsive phase and also near the time of the estimated halo cme launch time .
radio pulsations can be due to magnetic reconnection , and recently have associated this with plasmoid formation .
the projected radio sources of this emission were above the active region , which is where the halo cme most probably originated .
the source heights for the fluctuations are uncertain , since the observed plasma emission could also reflect the density of the dense active region loops , along which beams of particles were travelling . as the bursts within the envelope had both positive and negative frequency drifts , it is possible that magnetic trapping was present .
the drifting radio feature started next , at 13:14 ut near 500 mhz , which suggests a lower height in the atmosphere compared to the earlier narrow band fluctuations .
the estimated speed for the burst driver is difficult to calculate because of the high starting frequency ( high frequencies and high densities require that the standard atmospheric models have to be multiplied by arbitrary adjustment factors ) . from the frequency drift and the density scale height
we can make a rough estimate that the speed was around 1420 km s@xmath13 . in the radio images the projected source speed was similar
, 12001400 km s@xmath13 .
the source trajectory suggests that this feature could have been associated with the eit wave .
the eit wave features appeared near the locations of the drifting burst sources at 13:17 ut , near the bright euv `` rim '' shown in figures [ fig:7 ] and [ fig:9 ] .
unfortunately we could not calculate the speed of the eit wave , as it was observed very near the solar limb .
the halo cme speed was much higher than the estimated speed of this radio source , and the main halo - cme loop was also moving more to the west compared to the motion of the radio source .
the radio source could also be showing a filament trajectory , as the observed euv ribbons indicated a filament eruption .
.summary of observations related to the 9 november 2002 event [ cols="<,<,<",options="header " , ] the dh type ii burst that we were able to follow from 14 mhz down to 2 mhz , after which it was still observed as patchy emission down to kilometer waves , looks to have a clear association with the cme front .
the estimated burst heights fit well with a cme bow shock scenario .
the deduced type ii burst driver speed at 13:2313:47 ut ( @xmath02070 km s@xmath13 ) is not too far off from the plane - of - the - sky cme leading front speed measured at 13:3113:42 ut ( 1895 km s@xmath13 ; acceleration to 2000 km s@xmath13 near 30 r@xmath12 ) or the measured cme lateral expansion speed , @xmath02300 km s@xmath13 before 14:00 ut .
the height time profiles of the type ii burst and the cme suggest that both were formed during the flare impulsive phase .
a timing match between the flare impulsive phase and the cme acceleration phase have been noted by , _
e.g. _ , . the onset time analysis for protons gives the most probable acceleration time around 13:5714:27 ut .
the best candidate for the accelerator is the propagating cme bow shock , visible as the ip type ii burst . near 14:00 ut
when the protons could have been released , the ( projected ) white - light cme front was located at height @xmath09 r@xmath30 , and the type ii burst height was approximately 79 r@xmath30 ( actual number depending on the density model used ) .
we emphasize that between 14:06 and 14:40 ut no other radio signatures were observed in the dynamic spectra .
the two type iii burst groups that were observed to start at metric wavelengths ( at 60 and 250 mhz ) and continued to ip space suggest that electrons were accelerated at low heliocentric heights where they had access to open field lines .
low acceleration heights have been reported earlier by , and their results of acceleration heights were between 0.1 and 0.5 r@xmath12 above the photosphere agree with our estimates .
they claim that acceleration usually happens well after the flare maximum .
the velocity dispersion analysis for the electron data , using the low path length obtained from the proton analysis , gives an estimate for the time of electron release at 13:4014:03 ut .
this time range agrees with our observations of the type iii burst group observed to start at 13:48 ut .
the start of the next type iii burst group ( 14:02 ut ) is somewhat late compared with the estimated electron injection time , but it is still possible .
acceleration happens well after the flare maximum ( 13:23 ut ) in this case .
but , if we accept the high path length obtained from the analysis of electron data only , the release time ( 13:2313:39 ut ) will match with the flare maximum .
have suggested that accelerated electrons can be injected into the earth - connected flux tube when a coronal transient ( eit wave ) hits the field lines . in our case
we see the eit wave passing the southwest `` rim '' approximately around 13:2513:48 ut , but there is uncertainty due to the poor eit image cadence .
this time frame agrees with the estimated electron injection times with both the low and high path lengths and is close to the start time of the first type iii burst group .
however , since the origin of euv emission is relatively low in the solar atmosphere compared to the observed radio emission of the first two type iii burst groups , the association remains questionable .
one lasco image shows that the halo cme had a brightening on the side where the earlier , slower cme was propagating .
this brightening appears between the available lasco c2 images at 13:31 and 13:54 ut .
since the latitudinal and longitudinal distance between the two cme source regions was only 2040@xmath1 , it is possible that the halo cme interacted with the earlier cme or with a streamer located in the northwest .
a shock front could have formed in the interaction region , due to compression .
the radio enhancement was observed during 13:3013:40 ut at 3565 mhz , where the corresponding heliocentric height is around 1.52.0 r@xmath12 . at 13:54 ut the white - light brightening on the cme side had a height of @xmath04 r@xmath12 .
the deduced speed of the disturbance , if the radio feature and the white - light brigtening were connected , is approximately 11001500 km s@xmath13 .
we also observed tilted type iii burst lanes at times when the type iiis passed the ip type ii burst .
after the tilt , some of the type iii burst lanes brightened .
we present in figure [ fig:13 ] one possible geometry leading to the observed dynamics .
our sketch also explains the observed bidirectional motion of the metric type iii bursts : + electrons are accelerated in the possible reconnection site a2 in the low corona , where they have access to open magnetic field lines ( some open field lines were present in the potential magnetic field maps prior to the event ) .
the electron beams stream both upwards and downwards , as observed in the dynamic spectra and verified in nrh radio images .
site a1 is also a possible reconnection site , because of the open field lines near the active region that produced the earlier cme1/earlier observed streamer region .
since a1 site is located higher in the solar atmosphere than site a2 , a1 could be the acceleration region for the second type iii burst group ( start frequency near 60 mhz ) and site a2 could be the the acceleration region for the later birectional type iii burst group ( start frequency near 250 mhz ) .
+ accelerated energetic electrons then enter the compression region b which is behind the shock front driven by cme2 .
the result is a bending of the type iii burst lanes as the density gradient along the line of force is diminished .
particles propagate ahead of the shock wave , to region c , and experience a sudden decrease of the local density and magnetic field .
this in turn results in a drop of plasma frequency and focusing of the electron beam .
the latter , in turn , leads to a more efficient generation of langmuir waves and , hence , a re - brightening of the radio burst upstream of the shock front .
the release times for solar energetic particles , electrons and protons , can be obtained from , _
e.g. _ , velocity dispersion analysis .
the feasibility of this method can be argued , as the path lengths can vary and there can be unknown propagation effects in the ip space . in our analysis protons in the 1480 mev energy range yield path lengths of 1.01.4 au , while the analysed electrons suggest much higher path lengths of 2.42.7 au .
the path lengths of course have a strong effect on the estimated release times : with the low path lengths , electrons would be released just before the protons , and with adopting the high path lengths just for the electrons , electron release would have preceded proton release by more than half an hour . very high path lengths are usually taken as an indication of not - so - accurate timing results ( lintunen and vainio , @xcite ; saiz _ et al .
_ , @xcite ) .
high path lengths can be the result of delayed onset observation on low - energy channels due to high pre - event backgrounds , high contrast between the spectral indices of pre - event background and the sep event , prolonged injection of the particles , or interplanetary scattering of the particles . therefore the release times yielded by the lower path lengths are more plausible .
we do not speculate on the idea of a `` seed population '' for the released particles , although there is the possibility of injecting ( flare ) particles into the corona where the particles get accelerated later .
the timing and the high start frequencies of the metric type iii bursts that continue to the ip space suggest , however , that at least the energetic electrons are accelerated low in the corona and after the flare maximum . in summary , from the estimated release times for the electrons and protons compared with the multi - wavelength signatures of flare and coronal mass ejection processes ,
we conclude that electrons were most probably accelerated at low coronal heights in the wake of a fast halo - type cme the source region could be reconnecting structures either on the side of the cme or near its legs and that the best candidate for proton accelerator is the cme bow shock , at a time when the cme was propagating high in the corona , above 9 r@xmath30 . the tilts and brightenings in the type iii bursts at the times when they pass the ip type ii burst is a new observation .
we have presented a possible geometry leading to the observed dynamics , and point out that the tilted type iii bursts can reveal propagating shock fronts in cases when the type ii burst lanes are not visible in the spectrum .
we thank the radio group at lesia , observatoire de paris , france , for the use of nanay radioheliograph and decameter array data , and the soho / ephin instrument team and the _ wind_/3dp instrument team and the coordinated data analysis web ( cdaweb ) for providing the particle data .
klein and r. gmez - herrero are thanked for fruitful discussions .
the comments and suggestions made by the anonymous referees improved the paper significantly .
the lasco cme catalog is generated and maintained by nasa and catholic university of america in co - operation with the naval research laboratory .
soho is an international co - operation project between esa and nasa .
n.j.l . wishes to thank the visl foundation of the finnish academy of science and letters for a travel grant to spm-11 .
the academy of finland is thanked for partial support under project n : o 104329 .
bougeret , j .- l . ,
kaiser , m.l . ,
kellogg , p.j . , manning , r. , goetz , k. , monson , s.j .
et al . _ : 1995 , _ space sci .
_ * 71 * , 231 .
brueckner , g.e . ,
howard , r.a . , koomen , m.j . , korendyke , c.m . ,
michels , d.j . , moses , j.d . , _
et al . _ : 1995 , _ solar phys . _ * 162 * , 357 .
cane , h.v . ,
erickson , w.c .
: 2005 , _ astrophys .
j. _ * 623 * , 1180 .
cane , h.v . , reames , d.v . , von rosenvinge , t.t . : 1988 , _ j. geophys
. res . _ * 93 * , 9555 .
cliver , e.w . ,
webb , d.f . , howard , r.a . : 1999 , _ solar phys _ * 187 * , 89 .
dalla , s. , balogh , a. , krucker , s. , posner , a. , mller - mellin , r. , anglin , j. d. , _ et al .
_ : 2003 , _ ann .
geophys . _
* 21 * , 1367 .
debrunner , h. , flueckiger , e.o . , lockwood , j. a. : 1990 , _ astrophys .
ser . _ * 73 * , 259 .
delaboudinire , j .-
p . , artzner , g.e . ,
brunaud , j. , gabriel , a.h .
, hochedez , j.f . , millier , f. , _ et al . _ : 1995 , _ solar phys .
_ * 162 * , 291 .
dulk , g.a . : 1985 , _ ann .
astrophys . _
* 23 * , 169 .
gopalswamy , n. , yashiro , s. , michaek , g. , kaiser , m.l . , howard , r.a . ,
reames , d.v . , leske , r. , von rosenvinge , t. : 2002 , _ astrophys . j. _ * 572 * , 103 .
gosling , j.t . : 1997 , in : crooker , n. , joselyn , j.a . ,
feynman , j. ( eds . ) , _ coronal mass ejections _
, geophysical monograph * 99 * , american geophysical union , washington , 9 .
hilchenbach , m. , sierks , h. , klecker , b. , bamert , k. , kallenbach , r. : 2003 , in : velli , m. , bruno , r. , malara , f. , bucci , b. ( eds . ) , _ solar wind ten _ , aip conf . proc . * 679 * , american institute of physics , new york , 106 .
huttunen - heikinmaa , k. , valtonen , e. , laitinen , t. : 2005 , _ astron .
astrophys . _ * 442 * , 673 .
de jager , c. : 1986 , _ space sci . rev . _ * 44 * , 43 .
kahler , s : 1994 , _ astrophys .
j. _ * 428 * , 837 .
kahler , s.w . : 2003 , _ adv .
space res . _
* 32 * , 12 , 2587 .
kallenrode , m. : 1993 , _
j. geophys .
res . _ * 98 * , 5573 .
kerdraon , a. , delouis , j. : 1997 , in : trottet , g. ( ed . ) , _ coronal physics from radio and space observations _ , lecture notes in physics * 483 * , springer - verlag , berlin heidelberg , 192 .
klein , k .-
chupp , e.l . ,
trottet , g. , magun , a. , dunphy , p.p . , rieger , e. , urpo , s. : 1999 , _ astron .
_ * 348 * , 271 .
klein , k .-
posner , a. : 2005 , _ astron .
astrophys . _ * 438 * , 1029 .
klein , k .-
trottet , g. : 2001 , _ space sci .
rev . _ * 95 * , 215 .
kliem , b. , karlicky , m. , benz , a.o . : 2000 , _ astron .
astrophys . _
* 360 * , 715 .
kocharov , l. , torsti , j. : 2002 , _ solar phys . _ * 207 * , 149 .
kontogeorgos , a. , tsitsipis , p. , caroubalos , c. , moussas , x. , preka - papadema , p. , hilaris , a. , petoussis , v. , bouratzis , c. , bougeret , j .- l . , alissandrakis , c.e . ,
dumas , g. : 2006 , _ exp .
_ * 21 * , 41 .
krucker , s. , larson , d.e . ,
lin , r.p .
, thompson , b.j . : 1999 , _ astrophys .
j. _ * 519 * , 864 .
krucker , s. , lin , r.p .
: 2000 , _ astrophys .
j. _ * 542 * , l61 .
leblanc , y. , dulk , g.a .
, bougeret , j - l . : 1998 , _ solar phys . _
* 183 * , 165 .
leblanc , y. , dulk , g.a . ,
vourlidas , a. , bougeret , j - l . : 2001 , _ j. geophys .
res . _ * 106 * , 25301 .
lecacheux , a. , : 2000 , in : stone , r.g . ,
weiler , k.w . ,
goldstein , m.l .
, bougerot j .-
( eds . ) , _ radio astronomy at long wavelengths _ , agu monograph series * 119 * , american geophysical union , washington , 321 .
lin , r.p . , anderson , k.a .
, ashford , s. , carlson , c. , curtis , d. , ergun , r. , _ et al . _ : 1995 , _ space sci .
_ * 71 * , 125 .
lin , r.p . , dennis , b.r . ,
hurford , g.j . ,
smith , d.m . , zehnder , a. , harvey , p.r .
et al . _ : 2002 , _ solar phys . _
* 210 * , 3 .
lintunen j. , vainio r. : 2004 , _ astron .
_ * 420 * , 343 .
mann , g. , klassen , a. , aurass , h. , classen , h - t . : 2003 , _ astron .
astrophys . _ * 400 * , 329 .
messmer , p. , benz .
monstein , c. : 1999 , _ solar phys . _ * 187 * , 335 .
mewaldt , r. , cohen , c. , haggerty , d. , gold , r.e . ,
krimigis , s.m . ,
leske , r.a . , _
et al . _ : 2003 , in : kajita , t. , asaoka , y. , kawachi , a. , matsubara , y. , sasaki , m. ( eds . ) , _ proc .
28th int . cosmic ray conference _ * 6 * , universal academy press , trukuba , 3313 .
m " uller - mellin , r. , kunow , h. , fleissner , v. , pehlke , e. , rode , e. , roschmann , n. , _ et al . _ : 1995 , _ solar phys .
_ * 162 * , 483 .
nelson , g.j . ,
melrose , d.b . : 1985 , in : mclean , d.j . ,
labrum , n.r .
( eds . ) , _ solar radio physics _
, cambridge univ . press ,
cambridge , 333 .
newkirk , g. jr .
: 1961 , _ astrophys .
j. _ * 133 * , 983 .
pohjolainen , s. , van driel - gesztelyi , l. , culhane , j.l . ,
manoharan , p.k . ,
elliott , h.a . : 2007 , _ solar phys . _ * 244 * , 167 .
reames , d.v . : 1999 , _ space sci .
_ * 90 * , 413 .
saito , k. : 1970 , _ ann .
. obs _ * 12 * , 51 .
siz , a. , evenson , p. , ruffolo , d. , bieber , j. w. : 2005 , _ astrophys . j. _ * 626 * , 1131 .
scherrer , p.h . ,
bogart , r.s . , bush , r.i . ,
hoeksema , j.t .
, kosovichev , a.g . , schou , j. , _ et al . _
: 1995 , _ solar phys .
_ * 162 * , 129 .
schrijver , c.j . ,
derosa , m.l .
: 2003 , _ solar phys .
_ * 212 * , 165 .
schwenn , r. , dal lago , a. , huttunen , e. , gonzalez , w.d .
: 2005 , _ ann .
geophys . _
* 23 * , 1033 .
simnett , g.m .
: 2005 , _ solar phys . _ * 229 * , 213 .
torsti , j. , anttila , a. , kocharov , l. , mkel , p. , riihonen e. , sahla , t. , _ et al .
_ : 1998 , _ geophys .
lett . _ * 25 * , 2525 .
torsti , j. , valtonen , e. , lumme , m. , peltonen , p. , eronen , t. , louhola , m. , _ et al .
_ : 1995 , _ solar phys . _ * 162 * , 505 .
tylka , a.j . ,
cohen , c.m.s . ,
dietrich , w.f .
, krucker , s. , mcguire , r.e . , mewaldt , r.a . , _
et al . _ : 2003 , in : kajita , t. , asaoka , y. , kawachi , a. , matsubara , y. , sasaki , m. ( eds . ) , _ proc .
28th int . cosmic ray conference _ * 6 * , universal academy press , trukuba , 3305 .
vilmer , n. , pick , m. , schwenn , r. , ballatore , p. , villain , j.p . : 2003 , _ ann .
_ * 21 * , 847 .
vrnak , b. , magdaleni , j. , zlopec , p. : 2004
, _ astron .
astrophys . _ * 413 * , 753 .
wang , y. , zheng , h. , wang , s. , pe , y. : 2005 , _ astron .
astrophys . _
* 434 * , 309 .
wild , j. p. : 1950
res . a _ * 3 * , 541 .
zhang , j. , dere , k.p . ,
howard , r.a . ,
kundu , m.r . , white , s.m . : 2001 , _ astrophys .
j. _ * 559 * , 452 . | a high - speed halo - type coronal mass ejection ( cme ) , associated with a goes m4.6 soft x - ray flare in noaa ar 0180 at s12w29 and an eit wave and dimming , occurred on 9 november 2002 .
a complex radio event was observed during the same period .
it included narrow - band fluctuations and frequency - drifting features in the metric wavelength range , type iii burst groups at metric hectometric wavelengths , and an interplanetary type ii radio burst , which was visible in the dynamic radio spectrum below 14 mhz .
to study the association of the recorded solar energetic particle ( sep ) populations with the propagating cme and flaring , we perform a multi - wavelength analysis using radio spectral and imaging observations combined with white - light , euv , hard x - ray , and magnetogram data .
velocity dispersion analysis of the particle distributions ( soho and _ wind _ _ in situ _
observations ) provides estimates for the release times of electrons and protons .
our analysis indicates that proton acceleration was delayed compared to the electrons .
the dynamics of the interplanetary type ii burst identify the burst source as a bow shock created by the fast cme .
the type iii burst groups , with start times close to the estimated electron release times , trace electron beams travelling along open field lines into the interplanetary space .
the type iii bursts seem to encounter a steep density gradient as they overtake the type ii shock front , resulting in an abrupt change in the frequency drift rate of the type iii burst emission .
our study presents evidence in support of a scenario in which electrons are accelerated low in the corona behind the cme shock front , while protons are accelerated later , possibly at the cme bow shock high in the corona . |
to determine the density dependence of the nuclear symmetry energy @xmath1 , thus the equation of state ( eos ) of neutron - rich nuclear matter , has been a longstanding goal of both nuclear physics and astrophysics .
the @xmath1 is critical for understanding not only the structure of rare isotopes and heavy - ion reactions @xcite , but also many interesting issues in astrophysics @xcite . in this contribution we first summarize some recent progress in constraining the @xmath1 from analyzing the isospin diffusion and @xmath0 ratio in heavy - ion collisions within an isospin and momentum dependent transport model ibuu04 @xcite .
we will then discuss some astrophysical ramifications of the partially constrained @xmath1 .
some interesting information about the @xmath1 has been obtained over the last few years from heavy - ion collisions @xcite . as an example , shown in the middle of fig .
[ esym00 ] are the @xmath1 used in the ibuu04 transport model .
the three mdi @xmath1 @xcite are obtained using a modified gogny force by adjusting the parameter @xmath2 introduced in the interaction . for comparisons , the iqmd @xcite and the apr@xcite predictions
are also shown . shown in the left window
is the degree of isospin diffusion @xmath3 @xcite versus the slope parameter @xmath4 of the mdi symmetry energy . within the ibuu04 model analysis , the msu data of tsang et al .
@xcite favors a @xmath1 between that with @xmath5 and @xmath6 around the apr prediction .
more quantitatively , one can infer that the slope parameter is about @xmath7 mev @xcite .
we notice that a very recent re - analysis of the msu isospin diffusion data using the imqmd model found that the extracted range of the @xmath1 overlaps with that from the ibuu04 analysis around the @xmath1 with @xmath5 ( or apr ) @xcite . in the msu isospin diffusion reactions the maximum density reached is about @xmath8 @xcite .
moreover , the isospin diffusion was found to probe the symmetry energy during the expansion phase of the reaction @xcite .
thus the isospin diffusion data provides us information about the @xmath1 in the sub - saturation density region . very interestingly ,
a recent ibuu04 transport model analysis of the fopi / gsi @xmath0 ratio data @xcite in relativistic heavy - ion collisions at sis / gsi with beam energies above 400 mev / a provides circumstantial evidence suggesting a rather soft nuclear symmetry energy at @xmath9 compared to the apr prediction @xcite . shown in the right window are the calculated @xmath0 ratios in comparison with the fopi data at 0.4 agev with the reduced impact parameter @xmath10 as a function of the neutron / proton ratio of the reaction system .
the inset shows the @xmath0 ratio as a function of @xmath11 for the @xmath12ru+@xmath12ru reaction at 0.4 agev .
it is seen that both the data and the calculations exhibit very weak @xmath11 dependence for the @xmath0 ratio , even for mid - central reactions where we found that the multiplicities of both @xmath13 and @xmath14 vary appreciably with the @xmath11 . for
the symmetric @xmath15ca+@xmath15ca and the slightly asymmetric @xmath12ru+@xmath12ru reactions , calculations using both @xmath16 and @xmath5 can well reproduce the fopi data .
most interestingly , for the more neutron - rich reactions of @xmath12zr+@xmath12zr and @xmath17au+@xmath17au calculations with @xmath18 and @xmath19 are clearly separated from each other .
the fopi data favors clearly the calculation with @xmath16 .
the corresponding symmetry energy is much softer than the apr prediction .
a detailed study on the excitation function of the @xmath0 ratio also indicates that the fopi data can be approximately reproduced only with the @xmath16 symmetry energy @xcite .
it is also interesting to mention that iqmd calculations @xcite give similar results as the ibuu04 with @xmath5 .
this is not surprising since the symmetry energy functionals used in the iqmd and the ibuu04 with x=0 are very similar for @xmath20 as shown in the middle of fig .
[ esym00 ] .
the most importance influence of a rather soft symmetry energy at supra - saturation densities , such as that with @xmath16 , on pion production is through the rather high neutron / proton ratio reached in the participant region due to the dynamical isospin fractionation . as shown in the middle window , with @xmath16
the @xmath1 at @xmath9 reached in the reaction is very small .
thus a rather high @xmath21 in the participant region is energetically favored due to the dynamical isospin fractionation @xcite and thus the larger @xmath0 ratio is observed . in the reactions considered at 400 mev / a
, the maximum central density reached is about @xmath22 . by varying separately the @xmath1 at sub- and supra - saturation densities used in the ibuu04 simulations for these reactions
it was found that the @xmath0 ratio is much more sensitive to the variation of the @xmath1 at supra - saturation rather than sub - saturation densities .
putting together the information from analyzing both the isospin diffusion and the @xmath0 ratio data , we expect that the @xmath1 reaches a maximum somewhere between @xmath23 and @xmath24 before it starts decreasing at higher densities .
this indicates the importance of mapping out the @xmath1 at densities from about @xmath23 to @xmath22 .
such experiments are being planned at several facilities .
if the @xmath1 is confirmed by more experimental and theoretical studies to decrease with increasing density above certain density , it not only posts a serious challenge to some nuclear many - body theories but also has important implications on several critical issues in nuclear astrophysics , such as , the cooling of proto - neutron stars , the possible formation of polarons due to the isospin separation instability @xcite , the possible formation of quark droplets @xcite and hyperons @xcite in the core of neutron stars .
since the @xmath1 is only partially constrained in some density regions by the available experimental data within the ibuu04 transport model analyses , astrophysical applications of these constraints involve some interpolations and/or extrapolations under some assumptions .
we notice here , however , that the extrapolation of any low - density symmetry energy to supra - saturation densities can be very dangerous as illustrated in fig .
1 . using the same ibuu04 transport model , at sub - saturation densities the isospin diffusion analyses favors clearly the apr prediction , but above about @xmath24 the @xmath0 ratio data favors instead the gogny prediction that is much softer than the apr prediction .
fortunately , while many astrophysical questions depend on the symmetry energy over the whole density range , some issues , such as the core - crust transition density in neutron stars depends only on the @xmath1 at sub - saturation densities . in the following section , we present several examples illustrating the astrophysical importance of the @xmath1 using the mdi eos with @xmath5 and @xmath6 extrapolated to supra - saturation densities .
effects of the softening of the @xmath1 , such as that with @xmath16 , is currently under investigation .
neutron stars are expected to have a solid inner crust surrounding a liquid core .
knowledge on properties of the crust plays an important role in understanding many astrophysical observations @xcite .
the inner crust spans the region from the neutron drip - out point to the inner edge separating the solid crust from the homogeneous liquid core . while the neutron drip - out density @xmath25 is relatively well determined to be about @xmath26 g/@xmath27 @xcite , the transition density @xmath28 at the inner edge is still largely uncertain mainly because of our very limited knowledge on the eos , especially the density dependence of the symmetry energy , of neutron - rich nucleonic matter @xcite
recently , using the equation of state for neutron - rich nuclear matter constrained by the isospin diffusion data from heavy - ion reactions in the same sub - saturation density range as the neutron star crust , xu et al .
put a tight constraint on the location of the inner edge of neutron star crusts @xcite .
the inner edge corresponds to the phase transition from the homogeneous matter at high densities to the inhomogeneous matter at low densities . in principle , the inner edge can be located by comparing in detail relevant properties of the nonuniform solid crust and the uniform liquid core mainly consisting of neutrons , protons and electrons ( @xmath29 matter ) .
however , this is practically very difficult since the inner crust may contain nuclei having very complicated geometries , usually known as the ` nuclear pasta ' @xcite .
furthermore , the core - crust transition is thought to be a very weak first - order phase transition and model calculations lead to very small density discontinuities at the transition @xcite . in practice , therefore , a good approximation is to search for the density at which the uniform liquid first becomes unstable against small amplitude density fluctuations with clusterization .
this approximation has been shown to produce very small error for the actual core - crust transition density and it would yield the exact transition density for a second - order phase transition @xcite .
several such methods including the dynamical method @xcite , the thermodynamical method @xcite and the random phase approximation ( rpa ) @xcite have been applied extensively in the literature . as a function of @xmath30 ( left windows ) and @xmath31 ( right windows ) by using both the dynamical and thermodynamical methods with the full eos and its parabolic approximation .
right : the transition pressure @xmath32 as a function of @xmath33 and @xmath30 within the dynamical method with the full eos and its parabolic approximation .
the mdi ( upper windows ) and skyrme interactions ( lower windows ) are used .
taken from ref .
@xcite . ] as a function of @xmath30 ( left windows ) and @xmath31 ( right windows ) by using both the dynamical and thermodynamical methods with the full eos and its parabolic approximation .
right : the transition pressure @xmath32 as a function of @xmath33 and @xmath30 within the dynamical method with the full eos and its parabolic approximation .
the mdi ( upper windows ) and skyrme interactions ( lower windows ) are used .
taken from ref .
@xcite . ] shown in the left panel of fig .
[ rhotpt ] is the @xmath33 as a function of @xmath30 and @xmath34 using both the dynamical and thermodynamical methods with the full eos and its parabolic approximation ( pa ) from the mdi interaction with the varying @xmath2 parameter and 47 skyrme forces @xcite . with the full mdi eos
, it is clearly seen that the @xmath35 decreases almost linearly with increasing @xmath30 for both methods .
this feature is consistent with the rpa results @xcite and that found recently by oyamatsu et al .
the similar relation is also observed between the @xmath33 and @xmath31 due to the fact that @xmath31 always correlate with @xmath30 for a fixed energy density functional @xcite .
it is interesting to see that both the dynamical and thermodynamical methods give very similar results . on the other hand ,
surprisingly , the pa drastically changes the results , especially for stiffer symmetry energies ( larger @xmath30 values ) . also included in the left panel of fig .
[ rhotpt ] are the predictions by kubis using the pa of the mdi eos in the thermodynamical approach @xcite . the large error introduced by the pa is understandable since the @xmath36-stable @xmath29 matter is usually highly neutron - rich and the contribution from the higher order terms in @xmath37 is appreciable .
this is especially the case for the stiffer symmetry energy which generally leads to a more neutron - rich @xmath38 matter at subsaturation densities .
in addition , simply because of the energy curvatures involved in the stability conditions , the contributions from higher order terms in the eos are multiplied by a larger factor than the quadratic term .
these features agree with the early finding @xcite that the @xmath39 is very sensitive to the fine details of the nuclear eos . applying the experimentally constrained @xmath30 values of @xmath40 mev to
the @xmath41 correlation obtained using the full eos within the dynamical method shown in the left panel of fig .
[ rhotpt ] , one can conclude that the transition density is between @xmath42 @xmath43 and @xmath44 @xmath43 .
this constrained range is significantly below the fiducial value of @xmath45 @xmath43 often used in the literature and the estimate of @xmath46 made in ref .
@xcite within the thermodynamical approach using the parabolic approximation of the eos .
the pressure at the inner edge is an important quantity related directly with the crustal fraction of the moment of inertia which can be measurable indirectly from observations of pulsar glitches @xcite .
it is very instructive to quote the analytical estimation obtained by lattimer and prakash @xcite for the transition pressure @xmath47,\end{aligned}\ ] ] where @xmath48 is the incompressibility of symmetric nuclear matter at @xmath23 and @xmath49 is the isospin asymmetry at @xmath33 .
one can see that , besides the implicit dependence on the symmetry energy through the @xmath33 and @xmath49 , the @xmath32 also depends explicitly on the value and slope of the @xmath1 at @xmath33 .
thus the @xmath32 depends very sensitively on the @xmath1 .
shown in the right panel of fig .
[ rhotpt ] is the @xmath32 as a function of @xmath33 ( left windows ) and @xmath30 ( right windows ) by using the dynamical method with and without the parabolic approximation .
the results from eq .
( [ lp ] ) using the @xmath33 and @xmath50 corresponding to the full eos and its pa are also shown for comparisons .
it is interesting to see that the eq .
( [ lp ] ) predicts qualitatively the same but quantitatively slightly higher values compared to the original expressions for the pressure with or without the pa for both the thermodynamical and dynamical methods even though this formula was derived from the thermodynamical method using the pa .
this observation implies that the direct effect of using the full eos or its pa on the pressure is small although the pa may affect strongly the transition pressure @xmath32 by changing the transition density @xmath33 .
the @xmath32 essentially increases with the increasing @xmath33 in calculations using the full eos , but a complex relation between the @xmath32 and @xmath33 is obtained using the pa .
the large difference in @xmath32 is due to the strong pa effect on the @xmath33 .
moreover , the latter does not vary monotonically with @xmath30 for the pa as shown in the right panel of fig .
[ rhotpt ] . as a function of @xmath33 by using the dynamical method without parabolic approximation for both mdi interaction and shf calculations .
the shaded band represent the constraint from the isospin diffusion data .
taken from ref .
@xcite . ]
it is also interesting to examine the range of @xmath32 corresponding to the @xmath33 and @xmath30 constrained by the heavy - ion reaction data . in fig .
[ ptrhotdyn ] , we show the @xmath32 as a function of @xmath33 by using the dynamical method and the full eos for both the mdi ( solid line ) and the skyrme ( filled squares ) calculations . it is interesting to see that the mdi and skyrme interactions give generally quite consistent results .
corresponding to the @xmath33 constrained in between 0.040 @xmath43 and 0.065 @xmath43 , the @xmath51 is limited between @xmath52 mev/@xmath53 and @xmath54 mev/@xmath53 with the mdi interaction as indicated by the shaded area , which is significantly less than the fiducial value of @xmath55 mev/@xmath53 often used in the literature @xcite . as pointed out in a recent work by avancini et al @xcite
, the value of @xmath55 mev/@xmath53 may be too large for most mean - field calculations without the pa .
we note that among the @xmath56 skyrme interactions used here , the following @xmath57 interactions , i.e. , the skmp , sko , r@xmath58 , g@xmath58 , sv , ski2 , ski3 , and ski5 , are consistent with the constraints from heavy - ion reactions .
these results indicate that one may introduce a huge error by assuming _ a priori _ that the eos is parabolic for a given interaction in calculating the @xmath35 and @xmath51 . for a given neutron star of total mass @xmath59 and radius @xmath60 , what are the respective sizes of its core and crust ?
how do they depend on the stiffness of the symmetry energy ?
how do they depend on the neutron star mass @xmath59 ?
how does the thickness of neutron star crusts depend on @xmath30 while it is well known that the size of neutron skin in heavy nuclei increases with the increasing @xmath30 @xcite ?
these questions have been investigated recently in ref .
@xcite . , the crust thickness @xmath61 , the core radius @xmath62 as functions of @xmath30 at fixed total mass of @xmath63 , @xmath64 and @xmath65 , respectively .
right : the crustal fraction of neutron mass @xmath66 , the moment of inertia @xmath67 of the whole star and the crust contribution @xmath68 as a function of @xmath30 , at fixed total mass @xmath63 , @xmath64 and @xmath65 , respectively . taken from ref .
@xcite.,title="fig : " ] , the crust thickness @xmath61 , the core radius @xmath62 as functions of @xmath30 at fixed total mass of @xmath63 , @xmath64 and @xmath65 , respectively .
right : the crustal fraction of neutron mass @xmath66 , the moment of inertia @xmath67 of the whole star and the crust contribution @xmath68 as a function of @xmath30 , at fixed total mass @xmath63 , @xmath64 and @xmath65 , respectively . taken from ref .
@xcite.,title="fig : " ] show in left panel of fig .
[ rl ] are the core radius @xmath62 , the crust thickness @xmath61 and the total radius @xmath60 as functions of @xmath30 for a fixed total mass of @xmath63 , @xmath64 and @xmath65 , respectively .
it is seen that the @xmath62 increases almost linearly with increasing @xmath30 .
the @xmath62 also increases with the increasing mass at a fixed @xmath30 .
this is because the stiffer the symmetry energy is , the larger the contribution of the isospin asymmetric part of the pressure will be , which makes the @xmath62 larger .
moreover , the @xmath61 decreases with the increasing @xmath30 especially for light neutron stars , as the transition density decreases with the increasing @xmath30 . as the thickness of the crust @xmath61 and the core radius @xmath62
depend oppositely on @xmath30 , the total radius @xmath69 of the neutron star may show a complicated dependence on @xmath30 .
we stress here that this is the result of a competition between the repulsive nuclear pressure dominated by the symmetry energy contribution and the gravitational binding .
interestingly , it is often mentioned that the crust of neutron stars bears some analogy with the neutron - skin of heavy nuclei .
however , they show completely opposite dependences on the @xmath30 .
namely , the size of neutron - skin usually increases with the increasing @xmath30 as a result of the competition between the nuclear surface tension and the pressure difference of neutrons and protons @xcite , while the thickness of neutron star crusts decreases with the increasing @xmath30 as a result of the competition between the nuclear pressure and the gravitational binding . what is the crustal fraction @xmath66 of the total mass and how does it depend on the symmetry energy ?
since the mass is simply the integration of the energy density , one expects the @xmath66 and @xmath70 have very similar dependences on @xmath30 @xcite .
shown in the right window of fig .
[ rl ] ( right panel ) is the @xmath71 .
the fractional mass of the crust decreases with the increasing @xmath30 at a fixed total mass , and it decreases with the increasing total mass at a fixed value of @xmath30 .
the moment of inertia is determined by the distribution of the energy density . from the middle window
, it is seen that the total moment of inertia increases with the increasing mass at a fixed value of @xmath30 and increases with the increasing @xmath30 at a fixed total mass .
the dependence on @xmath30 is relatively weak especially for light neutron stars .
however , the crust contribution of the moment of inertia varies much more quickly with @xmath30 .
it decreases with the increasing neutron star mass at a fixed value of @xmath30 and decreases with the increasing @xmath30 at a fixed total mass .
these are all consistent with the behaviors of the fractional mass and size of the crust .
the crustal fraction of the moment of inertia @xmath72 is particularly interesting as it can be inferred from observations of pulsar glitches , the occasional disruptions of the otherwise extremely regular pulsations from magnetized , rotating neutron stars .
it can be expressed approximately as @xcite @xmath73^{-1},\end{aligned}\ ] ] where @xmath74 is the mass of baryons and @xmath75 .
this analytical formula has been verified by direct numerical calculations using both the full eos and its pa @xcite .
furthermore , it is indicated that there exists big differences for @xmath76 by comparing calculations using the full eos and its pa @xcite , again due to the corresponding differences in the transition density . as it was stressed in ref .
@xcite , the @xmath76 depends sensitively on the symmetry energy at sub - saturation densities through the @xmath32 and @xmath33 , but there is no explicit dependence upon the eos at higher - densities .
experimentally , the crustal fraction of the moment of inertia has been constrained as @xmath77 from studying the glitches of the vela pulsar @xcite . combining the observational constraint of @xmath78 with the upper bounds of @xmath79 @xmath43 and @xmath80 mev/@xmath53 inferred from heavy - ion reactions ,
we can obtain a minimum radius of @xmath81 km for the vela pulsar . according to this constraint ,
the radius of the vela pulsar is predicted to exceed @xmath82 km should it have a mass of @xmath83 .
we notice that a constraint of @xmath84 km for this pulsar has previously been derived in ref .
@xcite by using @xmath85 @xmath43 and @xmath86 mev/@xmath53 . however , the constraint obtained here using data from both the terrestrial laboratory experiments and astrophysical observations is significantly different and actually it is more stringent .
gravitational waves are tiny disturbances in space - time and are a fundamental , although not yet directly confirmed , prediction of general relativity . they can be triggered in cataclysmic events involving ( compact ) stars and/or black holes .
they could even have been produced during the very early universe , well before any stars had been formed , merely as a consequence of the dynamics and expansion of the universe . because gravity interacts extremely weakly with matter
, gravitational waves would carry a genuine picture of their sources and thus provide undisturbed information that no other messenger can deliver @xcite .
gravitational wave astrophysics would open an entirely new non - electromagnetic window making it possible to probe physics that is hidden or dark to current electromagnetic observations @xcite .
deformed pulsars and various oscillation modes of spherical neutron stars are among the possible sources of gravitational waves . in particular ,
the deformed pulsars are major candidates for sources of continuous gravitational waves in the frequency bandwidth of the existing ground - based laser interferometric detectors including the ligo @xcite and virgo ( e.g. ref .
@xcite ) . while the oscillations of neutron stars are mostly at frequencies much higher than the currently existing and planned gravitational wave observatories , their studies are of fundamental theoretical interest .
the strain - amplitude of gravitational waves from deformed pulsars depends on the star s quadrupole moment determined by the eos of neutron - rich nuclear matter . on the other hand , both the frequency and the decay rate of the fundamental oscillation modes
are determined also by the eos of neutron - rich nuclear matter . in the following we present several examples illustrating the imprints of the symmetry energy on gravitation waves .
more details can by found in refs .
@xcite .
the strain amplitude of gravitational waves at the earth s vicinity ( assuming an optimal orientation of the rotation axis with respect to the observer ) from deformed pulsars can be written as @xcite @xmath87 where @xmath88 is the neutron star rotational frequency , @xmath89 its principal moment of inertia , @xmath90 its equatorial ellipticity , and @xmath91 its distance to earth . the ellipticity is related to the neutron star maximum quadrupole moment ( with @xmath92 ) via @xcite @xmath93 where for _ slowly _ rotating ( and static ) neutron stars @xmath94 can be written as @xcite @xmath95 in the above expression @xmath96 is the breaking strain of the neutron star crust which is rather uncertain at present time and lies in the range @xmath97 $ ] @xcite . from eqs .
( [ eq.1 ] ) and ( [ eq.2 ] ) it is clear that @xmath98 does not depend on the moment of inertia @xmath89 , and that the total dependence upon the eos is carried by the quadrupole moment @xmath94 . thus eq . ( [ eq.1 ] ) can be rewritten as @xmath99 with @xmath100 . for slowly rotating neutron stars
lattimer and schutz @xcite derived the following empirical relation for the moment of inertia @xmath101\ ] ] using eq .
( [ eq.5 ] ) to calculate the neutron star moment of inertia and eq .
( [ eq.3 ] ) the corresponding quadrupole moment , the ellipticity @xmath102 can be readily computed ( via eq .
( [ eq.2 ] ) ) .
the results are shown in fig .
it is clearly seen that the eos , especially the symmetry energy , has a strong effects on all of these important quantities .
since the global properties of spinning neutron stars ( in particular the moment of inertia ) remain approximately constant for rotating configurations at frequencies up to @xmath103 @xcite , the above formalism can be readily employed to estimate the gravitational wave strain amplitude , provided one knows the exact rotational frequency and distance to earth , and that the frequency is relatively low ( below @xmath104 ) .
shown in the left window of fig .
[ gw2 ] is the gw strain amplitude , @xmath98 , as a function of frequency for several slowly rotating near - earth neutron stars . at higher rotational frequencies ,
relativistic rotational models have to be used @xcite .
the solid line represents the designed upper detection limit of ligo .
shown in the right window is the @xmath98 for the psr j1748 - 2446 rotating at 716z assuming it has an ellipticity of @xmath105 .
in both cases clear imprints of the symmetry energy are seen . in the framework of general relativity ,
gravitational radiation damps out the neutron star oscillations which leads to the frequency of the non - radial oscillations to become `` quasi - normal '' ( complex ) with a real part representing the actual frequency of the oscillation and an imaginary part representing the losses due to its damping @xcite .
the eigen - frequencies of the quasi - normal modes could be found by solving the equations which describe the non - radial perturbations of a static neutron star in general relativity . the critical input to solve the equation is the nuclear eos .
the so called w - mode associated with the space - time perturbation only exists in general relativity .
it is very important for astrophysical applications since it is related to the space - time curvature and exists for all relativistic stars , including black holes .
frequency ( left ) and decay time ( right ) of @xmath106-mode ( upper ) and @xmath107-modes ( lower ) as functions of the neutron star mass @xmath59 . taken from ref .
@xcite , title="fig : " ] frequency ( left ) and decay time ( right ) of @xmath106-mode ( upper ) and @xmath107-modes ( lower ) as functions of the neutron star mass @xmath59 . taken from ref .
@xcite , title="fig : " ] using the mdi eos , wen et al .
recently studied the axial w - modes @xcite .
figure [ wen1 ] displays both the frequency and damping time of the @xmath106- ( upper frame ) and @xmath107-modes ( lower frame ) respectively , as a function of the neutron star mass .
the results establish the relationship between the expected frequencies of the axial w - modes , for a given eos , and the stellar mass .
it is interesting to notice that there is a clear difference between the frequencies calculated with the mdi ( @xmath5 ) eos and those with the mdi ( @xmath6 ) eos .
since the major difference between these two cases is the density dependence of the nuclear symmetry energy , it is obvious that the symmetry energy has a clear imprint on the frequencies . the scaled frequency ( left ) and decay time ( right ) of @xmath106-mode ( upper ) and @xmath107-modes ( lower ) as functions of the neutron star compactness @xmath108 . taken from ref .
@xcite , title="fig : " ] the scaled frequency ( left ) and decay time ( right ) of @xmath106-mode ( upper ) and @xmath107-modes ( lower ) as functions of the neutron star compactness @xmath108 . taken from ref .
@xcite , title="fig : " ] shown in fig .
[ wen2 ] are the real ( upper panel ) and imaginary ( lower panel ) parts of the eigen - frequency of @xmath106 and @xmath107-modes scaled by the mass @xmath59 as a function of the compactness parameter @xmath108 , respectively .
these results suggest that the scaled eigen - frequency exhibits a universal behavior as a function of the compactness parameter independent of the eos used .
as discussed by andersson @xcite , benhar @xcite and tsui @xcite this finding could be used to constrain the frequency and damping time of gravitational waves .
this is very important for guiding the gravitational wave search provided the mass and radius of the prospective source ( neutron star ) are known . on the other hand , when the gravitational wave astronomy becomes a reality , namely if both the frequency and dumping time for a given neutron star are known this could provide information on the neutron star mass and radius . in figs .
[ wen2 ] the fitting curves representing the results of tsui et al . @xcite are also shown .
it is seen that the numerical results of wen et al .
are in good agreement with those of tsui et al .
in summary , important progress has been made in recent years in constraining the symmetry energy with heavy - ion collisions .
their implications in some astrophysical phenomena have been explored . nevertheless , the field is still at its beginning .
while a number of potentially useful probes of the @xmath1 have been proposed , available experimental data are mostly for reactions with stable beams @xcite . coming experiments with more neutron - rich nuclei at several advanced radioactive beam facilities
are expected to improve the situation dramatically .
thus , more exciting times are yet to come .
this work was supported in part by the us national science foundation under grant no .
phy-0652548 , phy-0757839 , phy-0758115 and phy-0457265 , the welch foundation under grant no .
a-1358 , the research corporation under award no .
7123 , the texas coordinating board of higher education award no .
003565 - 0004 - 2007 , the national natural science foundation of china under grant nos .
10575071 , 10675082 , 10874111 , and 10647116 , moe of china under project ncet-05 - 0392 , shanghai rising - star program under grant no .
06qa14024 , the srf for rocs , sem of china , the national basic research program of china ( 973 program ) under contract no .
2007cb815004 , and the young teachers training program of the chinese scholarship council under grant no . | heavy - ion reactions especially those induced by radioactive beams provide useful information about the density dependence of the nuclear symmetry energy , thus the equation of state of neutron - rich nuclear matter , relevant for many astrophysical studies .
the latest developments in constraining the symmetry energy at both sub- and supra - saturation densities from analyses of the isopsin diffusion and the @xmath0 ratio in heavy - ion collisions using the ibuu04 transport model are discussed .
astrophysical ramifications of the partially constrained symmetry energy on properties of neutron star crusts , gravitational waves emitted by deformed pulsars and the w - mode oscillations of neutron stars are presented briefly .
address = department of physics , texas a&m university - commerce , commerce , texas 75429 - 3011 , usa address = institute of theoretical physics , shanghai jiao tong university , shanghai 200240 , china address = cyclotron institute and physics department , texas a&m university , college station , texas 77843 - 3366 , usa address = department of physics , texas a&m university - commerce , commerce , texas 75429 - 3011 , usa , altaddress = department of physics , san diego state university , 5500 campanile drive , san diego ca 92182 - 1233 , usa address = department of physics , texas a&m university - commerce , commerce , texas 75429 - 3011 , usa , altaddress = department of physics , south china university of technology , guangzhou 510641 , china address = department of physics , texas a&m university - commerce , commerce , texas 75429 - 3011 , usa address = department of physics , tsinghua university , beijing 100084 , china address = institute of theoretical physics , shanghai jiao tong university , shanghai 200240 , china , altaddress = cyclotron institute and physics department , texas a&m university , college station , texas 77843 - 3366 , usa address = department of physics , texas a&m university - commerce , commerce , texas 75429 - 3011 , usa , altaddress = institute of modern physics , chinese academy of sciences , lanzhou 730000 , china address = department of physics , tsinghua university , beijing 100084 , china |
fe , and to a lesser extent ni , are the most abundant elements in the earth s core and mantle .
@xcite their alloys and minerals are important in several disciplines , including materials science and engineering in particular .
fe - ni alloys show a rich structural phase diagram as a function of composition and temperature and a pronounced magneto - structural coupling .
the latter is most evident in the properties of the invar alloy ( at near 35 at.% ni ) , which has a near zero thermal expansion coefficient due to magnetic interactions .
@xcite another example of the interplay between structure and magnetism is the divergence of the magnetic spin reorientation transition thickness in fe@xmath4ni@xmath5 thin films on w(110 ) with @xmath6 approaching 0.3 .
@xcite interestingly , the phase diagram @xcite shows a miscibility gap at just around the same composition range , in which the alloy has a tendency to separate into fcc and bcc phases with different compositions .
therefore , understanding the phase decomposition behaviour of fe - ni alloy films on w(110 ) is important in order to clarify the structural factors contributing to the peculiar magnetic spin reorientation transition ( srt ) of this system . in general
, it has been difficult to experimentally observe the structural phase separation in bulk fe@xmath4ni@xmath5 near @xmath7 into spatially distinct regions of different composition .
several months of aging at temperatures in the range @xmath8c results in a heterogeneity at a length scale of about 10 nm .
@xcite samples collected from meteorites with a natural treatment at high temperatures and astronomical times show the same phenomenon at the macroscopic scale .
@xcite to speed up the phase separation process , vacancies were introduced via ion bombardment in order to enhance diffusion and rearrangement in the bulk , @xcite with direct consequences for the magnetic properties .
@xcite even in the absence of such conditions that favor phase separation , the presence of chemical inhomogeneities in the invar alloy after annealing and a subsequent increase in the curie temperature have been reported .
@xcite similarly , the in - plane to in - plane srt in fe@xmath4ni@xmath5 thin films on w(110 ) has been previously discussed in the context of a homogeneous strain within the fe lattice induced by randomly distributed ni atoms , @xcite which does not give any explanation for the observed divergence of the reorientation thickness at @xmath7 . along these lines , understanding the formation and energetics of heterogeneities in fe - ni alloys and alloy films is a prerequisite in discussing the physics involved both in the invar effect and in the fe - ni / w(110 ) srt . at the limit of ultrathin films , the presence and the nature of phase separation
was addressed in a recent work on a fe@xmath1ni@xmath2 monolayer on w(110 ) .
@xcite it was shown that the alloy monolayer decomposes into bcc ( pseudomorphic ) and fcc - like ( hexagonal ) monolayer regions .
the ni content in the bcc and fcc - like regions were measured to be @xmath9 and @xmath10 , respectively .
these compositions are similar to the bulk phase boundaries of fe - ni , though somewhat shifted towards higher fe content . nevertheless , the observation of the fcc - like hexagonal layer on the bcc(110 ) substrate even at the monolayer limit is important given that the srt in fe films on w(110 ) is reported to originate at the interface layer .
@xcite in this work , we focus on the fcc - bcc coexistence in thicker fe@xmath4ni@xmath5 ( @xmath7 ) films on w(110 ) at a thickness range closer to the spin reorientation transition .
we show that the phase separation occurs upon annealing a few - nanometers thick alloy film to about 300@xmath11c , and the micron - sized structural domains feature different compositions in agreement with the boundaries of the metastability region in the bulk phase diagram .
the observed crystal restructuring requires defect formation ( vacancy or interstitial ) , whereas the chemical heterogeneity calls for the diffusion of the alloy species .
the high interstitial formation energy above 3 ev for bcc fe @xcite makes it unlikely at the temperatures considered , which leaves vacancy formation as the more likely driving factor of the observed restructuring . based on this
, we focus our dft calculations on the vacancy formation energy in bcc fe@xmath12ni@xmath13 .
it will be demonstrated that the experimentally - determined energy barrier of [email protected] ev associated to the phase separation kinetics matches well with the calculations on the vacancy formation . in the following sections , we begin by describing the experimental details and theoretical methods .
we then present the experimental results on the phase separation in fe@xmath1ni@xmath2 films at high temperature and the chemical and structural characterization of the resulting heterogeneous surface .
the results of our dft calculations on the vacancy formation energy and surface diffusion barrier are reported in section iv , followed by a discussion of the energetics involved in the phase separation process .
the experiments were carried out with the spectroscopic photoemission and low - energy electron microscope ( speleem ) at the nanospectroscopy beamline ( elettra sincrotrone trieste ) .
@xcite the instrument combines low - energy electron microscopy ( leem ) with energy - filtered x - ray photoemission electron microscopy ( xpeem ) .
the former provides structural information , @xcite whereas the latter is used mainly to obtain chemical and magnetic distributions . @xcite the spatial resolution in the case of leem is better than 10 nm , whereas in xpeem the resolution is limited to slightly below 30 nm .
in addition to real - space imaging , the instrument can be used to acquire @xmath14-spot low energy electron diffraction ( @xmath14-leed ) data .
the transfer width in the diffraction mode is 13.4 nm .
the pressure in the microscope chamber was below @xmath15 mbar during growth and measurements .
the current study took advantage of the fast acquisition rates in leem to follow the growth and phase separation of the feni alloy films on w(110 ) in real time .
the coexisting phases were characterized structurally using dark - field leem and @xmath14-leed . in dark - field leem , the real - space image is obtained with an aperture placed in the diffraction plane that selects a particular leed beam and filters out all other electrons .
thus , the dark - field image reflects the distribution of the structural domains which gives rise to that diffraction spot . in the case
that the structural domains are large enough , @xmath14-leed is used to get the diffraction pattern from individual domains using an illumination aperture to limit the incident electron beam to a micron - sized spot on the surface .
the chemical heterogeneity was mapped using i ) 3p xps core - levels of fe and ni in energy - filtered xpeem mode and ii ) fe and ni l - edge xas by tuning the photon energy to the respective resonance and imaging the secondary photoelectrons . in the former
, the photon energy was varied from 250 ev to 650 ev in order to tune the inelastic mean free path of electrons and thus to evaluate the variation of the composition along the surface normal . in the latter , the probing depth is several nanometers due to the large mean free path of the secondary photoelectrons .
the w(110 ) substrate was cleaned with cycles of annealing in oxygen ( typically 1100 @xmath16c at @xmath17 mbar o@xmath18 for about 15 minutes in a preparation chamber attached to the microscope ) followed by high temperature flashes ( to above 2000 @xmath16c ) in uhv to remove surface oxygen .
the cleanliness was checked with leem and leed , which are very sensitive to the presence of contamination on the surface .
the fe@xmath4ni@xmath5 alloy film was grown on the w(110 ) substrate by codeposition of fe and ni from 2 mm thick high - purity rods installed in e - beam evaporators at a rate of about 0.25 ml / min .
the rate of each evaporant was determined within 5% by following the completion of the respective pseudomorphic monolayer on w(110 ) .
note that in the following sections the monolayer ( ml ) units will refer to a layer pseudomorphic to w(110 ) unless otherwise stated .
fe monolayer completion was monitored in leem at about 300 @xmath11c , at which the step flow growth was easily resolved .
the complete pseudomorphic ni monolayer was obtained above 100 @xmath11c . above a monolayer
, the ni layer transforms into @xmath19 and @xmath20 structures with successively higher packing densities .
@xcite the onset of the @xmath19 diffraction pattern in leed was used to calibrate the ni deposition rate . upon fe - ni codeposition ,
the composition of the resulting alloy was determined also from the ratio of the integrated 3p core - level photoemission signals of fe and ni , after correcting for the difference in photoionization cross - sections and in the microscope transmission which depend on the electron kinetic energy . in order to evaluate the vacancy formation energy and surface diffusion barrier
, we performed first - principles density - functional calculations as implemented in the pwscf code , @xcite which uses a plane - wave basis set and pseudopotentials .
exchange and correlation were described by the generalized - gradient approximation ( gga ) in perdew - burke - ernzerhof parametrization .
@xcite it is known that local - density approximation overstabilizes the nonmagnetic hcp and fcc phases of fe and that gga describes well the fe ground state .
@xcite the choice of gga gets further confirmation from the previous reports in which it is successfully applied to calculate the vacancy formation energy in bcc fe . @xcite we employed vanderbilt ultra - soft pseudopotentials generated using the following atomic configurations : @xmath21 for fe and @xmath22 for ni .
the nonlinear - core correction to the exchange and correlation potentials was included for both fe and ni .
our plane - wave basis kinetic energy cutoff was 41 ry for the wave - functions and 270 ry for the charge density .
a @xmath23 monkhorst - pack mesh was used for integrations over the brillouin zone .
a gaussian broadening of 0.01 ry was used for determination of the fermi energy .
the positions of the atoms in all calculations were relaxed until the residual forces on the atoms were smaller than @xmath24 ry / bohr .
for the lattice parameter , we used our optimized value of bcc bulk fe@xmath25ni@xmath26 , equal to 2.83 .
the monovacancy - formation energy in bcc @xmath27 was calculated at constant volume ( relaxing only the atomic positions in a supercell ) as an average of vacancy formation energies in @xmath28 and @xmath29 in order to circumvent the minor differences in energy due to the variation in the supercell stoichiometry when adding or substracting an atom . in this scheme , we started by evaluating the energy of the supercell containing 53 atoms , _
i.e. _ 36 fe and 17 ni atoms and a vacancy .
subsequently , we performed two bulk calculations with the vacancy filled . in one calculation ,
the vacancy was filled with a ni atom , and in the other with an fe atom , thus obtaining the @xmath30 and @xmath31 configurations , respectively . by averaging the energies for the two different reference configurations
the vacancy formation energy was obtained for composition @xmath32 ( @xmath27 ) .
the supercell size was chosen to be @xmath33 with 54 atoms , which had been shown to be sufficient for convergence in @xmath34-fe calculations @xcite .
such a supercell is optimal for the compositions around @xmath35 .
we constructed 3 different supercells , of which the first two feature random distributions of ni sites and the third one is an ordered @xmath28 alloy .
in the random supercells , the ni positions did not follow any specific rule .
in the ordered one , the ni sites were chosen randomly only in the first atomic layer , and the following layers were obtained by shifting the positions along the @xmath6-direction from one layer to the next . for the vacancy formation energy , we performed a total of 14 different calculations using three different supercells and 4 to 5 vacancy configurations for each one .
the calculation of the surface diffusion barriers was performed using a slab geometry .
the surfaces were modeled by `` asymmetric slabs '' in which at one termination the atomic positions were frozen at their bulk positions , and all other positions were atomically relaxed .
the adsorbate ni atom was held fixed laterally at a given adsorption site ( short and long bridge ; top ) , while its @xmath36-coordinate and the atoms at the surface and below were relaxed .
the periodic images of the slab were separated by a vacuum layer of 14 .
we used a @xmath37 unit cell with a ( 110 ) surface orientation and the slab consisted of 5 atomic layers .
fe@xmath4ni@xmath5 growth by codeposition of fe and ni on w(110 ) was carried out up to film thicknesses of 20 ml for various stoichiometries ranging from @xmath38 to @xmath39 .
for the case of @xmath40 , the leed data corresponding to the relevant stages during room temperature growth are displayed in fig .
[ fig : growth ] .
up to one monolayer coverage , the @xmath41 substrate pattern is preserved without any broadening of the diffraction spots . upon completion of the pseudomorphic fe@xmath1ni@xmath2 monolayer
, a @xmath19 superstructure appears , the spot intensity reaching a maximum at a total feni coverage of 1.22 ml @xcite as seen in the second panel in fig .
[ fig : growth ] .
interestingly , this behaviour is similar to that of a pure ni layer on w(110 ) @xcite in spite of the fe - rich alloy composition .
the @xmath20 structure observed for ni / w(110 ) at higher coverage @xcite is absent in the case of fe@xmath1ni@xmath2/w(110 ) . above 1.22 ml ,
the @xmath19 pattern continuously evolves into another superstructure , which fits well with the misfit dislocation structure observed for pure fe films on w(110 ) .
@xcite the misfit dislocation pattern ( third panel in fig .
[ fig : growth ] ) persists up to about 10 ml .
further growth at room temperature results in the ( @xmath42 ) bcc(110 ) pattern seen in the rightmost panel in fig .
[ fig : growth ] .
the position of the fractional spots indicates approximately 10% contraction with respect to the w(110 ) lattice , in good agreement with the fe(110 ) surface .
moreover , the broadening of the diffraction spots and the streaks along the @xmath43 $ ] direction are also observed for fe films on w(110 ) of comparable thickness .
the streaks contain weak diffuse facet spots that move in reciprocal space along @xmath43 $ ] as a function of energy ( a direct consequence of faceting ) , as in the case of fe / w(110 ) .
the surface of films prepared at room temperature has a slight granularity below the 50 nm length scale .
however , the presence of fcc domains is ruled out based on the leed and dark - field leem data .
moreover , xpeem imaging at the fe 3p and ni 3p core levels show no chemical heterogeneity above the resolution limit of about 30 nm . possibly , this roughness is associated with the nanoscale faceting of the room temperature grown surface as hinted in the elongated spot profiles in the rightmost panel in fig .
[ fig : growth ] . as we will discuss in the following sections , the fcc - bcc phase separation upon annealing takes place at a much longer length scale than the disorder present in the room temperature grown films . as illustrated in fig .
[ fig : growth ] , a room temperature codeposition of fe and ni up to a total thickness above 10 ml results in bcc(110 ) fe@xmath1ni@xmath2 films .
the disorder in the room - temperature grown films , visible in the broad diffraction beam profiles , smoothens out upon annealing the surface to temperatures below 250@xmath11c without changing the bcc order . at about 400@xmath11c and above
, the film breaks apart .
however , annealing to temperatures between 250@xmath11c to 400@xmath11c an additional leed pattern of hexagonal symmetry appears , accompanied by the formation of islands at the micron scale .
ni@xmath2 film in leem .
a ) 260@xmath11c , b ) 310@xmath11c , c ) 320@xmath11c , d ) 340@xmath11c , e ) 360@xmath11c , f ) 375@xmath11c .
electron energy is 19 ev . the temperature ramp from ( a ) to ( f )
spans a period of 1900 s.,width=302 ] the growth of fcc regions for an 18 ml fe@xmath1ni@xmath2 film can be followed in the leem sequence shown in fig .
[ fig : anneal ] . up to above 260@xmath11c
the surface remains homogeneous with a uniform change in electron reflectivity due to a progressive flattening / ordering of the surface .
following a slow temperature ramp ( about 10@xmath11c / min ) , triangular regions appear at about 300@xmath11c .
these grow slowly at low temperature and faster at higher temperature . experiments for different film stoichiometries
indicate that alloys with higher ni content tend towards a larger surface coverage of islands .
the island shapes are elongated in the bcc @xmath43 $ ] direction .
this is qualitatively consistent with the faster diffusion along @xmath43 $ ] .
however , the measured aspect ratio of the islands is near 2 , which describes a more anisotropic shape than what is expected from surface diffusion alone .
we tentatively attribute the island shape to ( at least in part ) the boundary energy between the fcc and bcc domains .
the leed pattern of such a heterogeneous surface is the superposition of two ( @xmath42 ) unit cells as shown in fig .
[ fig : characterize]a .
one of the two corresponds to the symmetry of a bcc(110 ) surface , with the lattice constant very close to that of a fe(110 ) surface . using the w substrate lattice as a reference ,
the lattice constant of this bcc pattern is very slightly smaller ( by 0.5% ) than that of fe(110 ) .
the other leed pattern represents a hexagonal structure and has a lattice vector expanded by 2.2% with respect to the ni fcc(111 ) unit cell .
dark - field leem images ( fig .
[ fig : characterize]b ) indicate that the hexagonal pattern originates from the islands whereas the rest of the film has the bcc(110 ) structure . in order to better understand the structure of the islands , the energy dependent electron reflectivity curves for the two regions
are plotted in fig .
[ fig : characterize]c .
the islands show a characteristic double peak i(v ) spectrum at around the first bragg energy , which is very similar to the i(v ) curve of the clean ni(111 ) surface .
@xcite therefore , based on the hexagonal kinematic leed pattern and the energy dependence of electron reflectivity , we conclude that the islands have the fcc structure . a careful inspection of the leem i(v ) data reveals that thickness - dependent intensity oscillations are not uniform over the surface .
[ fig : characterize]d displays the intensity derivative versus out - of - plane momentum transfer , in which the regular quantum oscillations are evident .
the film thickness is directly related to these oscillations , @xcite with thicker films resulting in a shorter period and a smaller amplitude due to reflectivity attenuation effects induced by the limited electron mean free path .
thus , the thickness can be laterally mapped from the local i(v ) spectra , as shown in fig .
[ fig : characterize]e .
the large fcc islands are thicker than the bcc regions by about 15% on average , although they also feature regions with reduced thickness .
the smaller islands instead are systematically thinner than the rest of the film .
importantly , the variations in the thickness point to a mechanism involving mass transport over distances above a micron .
moreover , the well - defined quantum oscillations in electron reflectivity show that the structure is homogeneous along the entire film depth both in the bcc and fcc regions .
the question arises to the chemical composition of the structurally heterogeneous surface as revealed by leem .
laterally resolved fe and ni distributions are displayed in the xas - peem images in fig .
[ fig : xpeem]a and b , respectively .
the chemical maps are given by the secondary photoelectron images obtained at photon energies around the l@xmath44 x - ray absorption threshold of the respective element .
more precisely , the image acquired with a photon energy at the center of resonance ( resonant image ) is normalized to that acquired at a photon energy several ev below the resonance ( baseline image ) .
in such a normalized xas - peem image , a value of unity corresponds to the absence of the respective element . using the average composition ( @xmath40 ) and the area coverage of the two phases as input
, the lateral map of the fe@xmath4ni@xmath5 alloy stoichiometry ( @xmath6 ) can be obtained from the x - ray intensity images , as seen in fig .
[ fig : xpeem]c . qualitatively , fig .
[ fig : xpeem]c shows that i ) there is a clear composition difference between the two phases , and ii ) the composition is nearly independent of the thickness variations .
quantitatively , we can assign @xmath45 and @xmath46 to the bcc and fcc phases , respectively .
the pronounced variation in the composition is a direct proof that the bcc - fcc phase separation is not due to a local or martensitic structural transformation but instead it involves considerable mass transport . beyond determining the compositions ,
the xpeem images show a qualitative difference between fe and ni distributions .
apart from the variations correlated to film thickness , the fe signal is rather uniform everywhere on the surface . instead , the fcc islands have more than a twice stronger ni signal compared to the bcc film .
we note that the xas data displayed in figs .
[ fig : xpeem]a and [ fig : xpeem]b reflect the total fe and ni amounts , respectively , along the film normal due to the relatively long mean free path of the secondary photoelectrons .
therefore , these observations suggest that the chemical heterogeneity is formed by ni mass transport laterally across the film , whereas fe mostly remains in place .
the fact that the fcc islands are thicker than the nominal film thickness confirm this statement .
however , the quantitative difference in thickness between the fcc and bcc regions is not sufficient to rule out lateral fe diffusion .
we have also acquired x - ray photoemission spectra at the fe 3p and ni 3p core levels for a different nominal alloy composition ( shown in the supplementary material ) .
importantly , the compositions extracted from the laterally - resolved xps data are dependent on the photon energy .
in particular , both phases appear more ni rich at energies with shorter electron inelastic mean free path , indicating that the surface is enriched in ni .
finally , we note that the fcc - bcc coexistence was observed in the entire composition range under study ( @xmath38 to @xmath47 ) .
qualitatively , the same behaviour was observed in all experiments .
importantly , at the highest ni content considered ( @xmath39 ) the fcc structure was present already after the room temperature growth and before annealing .
moreover , the length scale at which the fcc - bcc separation takes place was observed to be considerably shorter . at compositions below @xmath38 ,
the bcc phase was stable up to the dewetting temperature of the film , which is at around 400@xmath11c for the thicknesses considered .
the fcc island growth was monitored in real time using leem .
the result is given in fig .
[ fig : kinetics ] .
firstly , we observe that the island coverage tends to a certain value for the given nominal composition , and that there is a variation in the growth of individual islands .
the latter is evidenced in fig .
[ fig : kinetics]a , in which the growth of the fcc island labeled ` 2 ' seems impeded by the growth of the larger islands nearby .
this is consistent with material exchange over long distances , which causes a competition between nearby islands in capturing material .
[ fig : kinetics]b shows a plot of the fractional fcc island coverage as a function of time at constant temperature ( 340@xmath11c ) .
notably , the time dependence is reproduced reasonably well with a single exponential function .
similar experiments at different temperatures show that the time constant has an arrhenius behaviour .
in other words , the temperature - dependence of the dynamics suggests the presence of an energy barrier of @xmath48 ev .
the final fcc coverage is 25% for the data displayed in fig .
[ fig : kinetics ] .
other measurements qualitatively indicate that this value depends on the nominal composition , with higher ni content favoring the abundance of the fcc structure .
, the temperature is increased from below 200@xmath11c to 340@xmath11c and is kept constant afterwards .
the inset shows the time constant of the exponential fit vs inverse temperature obtained from a series of similar experiments at varying temperatures .
the small abrupt variations in the measured island coverage is due to readjustments of the objective focus , which slowly drifts at high temperature.,width=302 ] although the single exponential function approximately represents the entire time scale , the data in fig .
[ fig : kinetics]b show a faster rise immediately after nucleation and a longer tail at later stages of island growth .
this is an indication that there is more than one energy barrier contributing to the relaxation towards equilibrium .
therefore , we consider the more general case , in which there is a distribution of barriers .
the excellent fit in fig .
[ fig : kinetics]b shows that the time evolution can be represented by a stretched exponential function , @xcite which is often applied to relaxation in disordered systems .
the stretched exponential is described by a time dependence in the form @xmath49 .
the exponent @xmath50 is found to be 0.70 for the data displayed in fig .
[ fig : kinetics]b . the unitless quantity @xmath50 carries information on the distribution of energy barriers .
@xcite in particular , the width of the barrier distribution scales as @xmath51 . taking a gaussian barrier distribution , the full width at half maximum ( fwhm ) is found to be about 0.17 ev for @xmath52 . in order to give a physical sense to this energy barrier distribution , in the following section we will present the results of dft calculations for the energy parameters associated with the bcc @xmath35 alloy .
in figure [ fig : vac_en]a we show an example of the calculated vacancy formation energy ( vfe ) for the fe@xmath53ni@xmath54 alloy . it is given as a function of the chemical potential difference between the alloy and the elemental bulk phases , shown for fe and ni respectively on the bottom and top horizontal axes .
the maximum range of the chemical potential differences are derived from the calculated formation enthalpy , which is @xmath55 ev / atom for the particular supercell in figure [ fig : vac_en]a . as can be seen
, the vacancy formation energy depends only weakly on the chemical potential .
moreover , this dependence is identical for all the calculations performed ( see supplementary material ) .
therefore , for simplicity , in the following we will use the values of vacancy formation energies for the fe chemical potential equal to the bulk fe chemical potential . in order to account for a possible influence of the immediate environment of the vacancy
, we have performed calculations keeping track of the number of ni atoms near the vacancy . in particular , the stoichiometry and the randomness of the vacancy and ni positions are taken into account in dividing the total of 14 calculations according to the number of first ni neighbours into groups of 5 calculated configurations with 2 first ni neighbors , 4 with 3 , 3 with 4 , and 2 with a single first ni neighbor .
this choice roughly corresponds to the weights @xcite each number of neighbors has in a random fe@xmath53ni@xmath54 alloy .
note that we have not considered vacancy configurations with 5 to 8 ni neighbours , as they have a negligible contribution at this stoichiometry .
figure [ fig : vac_en]b presents a histogram of the resulting vacancy formation energies .
the histogram can be roughly described by a gaussian distribution centered around 1.71 ev with a fwhm deviation of 0.37 ev .
these values are in good agreement with the very recent results obtained using kinetic monte carlo calculations , which reported a range of vacancy formation energies between 1.2 ev and 2.15 ev for fe@xmath56ni@xmath56 .
@xcite figure [ fig : vac_en]c shows the vacancy formation energy as a function of the total number of ni neighbors in the first two shells . from the clear correlation , it can be inferred that the vacancy formation energy is reduced for configurations with increasing number of ni neighbors .
this is in line with our observation that the ni atoms , constrained in the bcc environment , relax significantly more than the fe atoms around the vacancy , and , in consequence , tend to reduce the vacancy formation energy .
we note that the strongest ni relaxations occur when more ni atoms are grouped together rather than sitting on scattered sites around the vacancy .
the tendency towards a value somewhat above 2 ev / vacancy in the absence of ni neighbours can be seen in figure [ fig : vac_en]c .
this is consistent with the theoretical vacancy formation energy of 2.2 ev in pure fe .
@xcite nevertheless , we also encounter differences in the vacancy energy for equivalent configurations of ni neighbours near the vacancy within different supercells .
we attribute such differences to long range elastic interactions , which depend on the entire supercell configuration .
lastly , we also calculate the diffusion barrier for a ni adatom on the bcc alloy surface .
we focus on the surface diffusion of ni adatoms because of the experimental observation that the surface is enriched in ni , as reported in the previous section .
the lowest energy path is found to be along the direction connecting the long - bridge site ( the global energy minimum ) and the short - bridge site , i.e. along the @xmath57 $ ] direction .
the diffusion barrier is calculated to be 0.34 ev .
this value compares well with the calculation of a fe / fe(110 ) diffusion barrier of 0.36 ev , @xcite and is considerably lower than the barrier of 0.65 ev for ni / ni(111 ) .
the experimental results reported in the preceding sections showed that there is a structural and chemical rearrangement at the micron scale of the fe@xmath4ni@xmath5 alloy film close to @xmath40 . vacancy formation and surface diffusion are the two important ingredients in the restructuring of the film .
it has been known that the surface of fe - ni alloys has a tendency to be enriched in ni .
@xcite by varying the probing depth via the inelastic mean free path of photoelectrons in photon energy dependent xps ( in addition to the xas spectra acquired with the secondary photoelectrons ) we found that ( i ) there is no surface segregation in the room - temperature grown films , ( ii ) upon annealing , the surface is enriched in ni both in the fcc and bcc phases . in short ,
we attribute the chemical heterogeneity predominantly to the surface diffusion of ni , with the calculated diffusion barrier of 0.34 ev as reported in the previous section .
the calculated average vacancy formation energy of 1.71 ev / vacancy is in good agreement with the experimental energy barrier of @xmath58 ev found from the phase separation kinetics .
moreover , the experimental kinetics indicates the presence of a barrier distribution with a fwhm of 0.17 ev .
this is also reproduced in the calculated distribution of vacancy formation energies , although with a larger spread of 0.37 ev .
therefore , considering the much lower surface diffusion barrier , we conclude that the rate limiting step in the phase separation in ultrathin alloy films is vacancy formation .
the difference between the theoretical vacancy formation energy and the experimentally evaluated barrier lies in the temperature dependence of the vacancy formation _
free _ energy .
it has been recently reported that there is a pronounced change in the energy parameters associated with mono - vacancies in bcc fe as a function of temperature .
@xcite in particular , the reported decrease in the vacancy formation free energy from 0 k to 600 k is about 10% which quantitatively agrees with the difference between our zero - temperature dft calculations and our experimental result at 600 k for the bcc fe - ni alloy . as opposed to time dependence ,
the length scale at which the phase separation takes place reflects the pronounced surface diffusion at around 300@xmath11c .
moreover , the fcc island shapes are elongated along the @xmath43 $ ] direction , which is the faster diffusion direction on the bcc(110 ) surface .
therefore , we attribute the mesoscopic morphology of the phase - separated film to surface diffusion , while the entire phase - separation and the atomic restructuring process is limited by vacancy formation .
importantly , these results clarify the main distinction between an alloy film ( or the alloy surface ) and a bulk fe - ni alloy .
although the surface appears crucial in determining the microscale morphology , interstitial diffusion barriers in the bcc fe and its alloys are just as low as the surface diffusion barriers .
@xcite instead , the interstitial formation energy is considerably larger than that of vacancies , and our study singles out vacancy formation as the limiting step in the phase separation process . thus , the principal role of the surface in facilitating the phase separation is the creation of vacancies without forming interstitials .
in summary , we have experimentally shown that the fe@xmath1ni@xmath2/w(110 ) film of thickness close to 20 atomic layers separates into micron - sized fcc and bcc regions at around 300@xmath11c .
the two structural regions have distinctly different compositions and the composition difference originates from a difference in the ni amount .
in addition , the energy barrier associated with the kinetics of phase separation matches the theoretically calculated vacancy formation energy .
moreover , the calculated vacancy formation energy is slightly different for each vacancy configuration , giving a distribution of energy barriers in agreement with experimental observation . in the phase separation process
, the surface facilitates the creation of vacancies , as well as helping ni diffusion with surface ni enrichment and the pronounced diffusion length .
we expect that the observed fcc - bcc phase separation in fe - ni thin films at the mesoscopic scale will provide a model for studying the effect of chemical and structural heterogeneity on the alloy magnetic properties near the invar composition . | the phase separation occurring in fe - ni thin films near the invar composition is studied by using high resolution spectromicroscopy techniques and density functional theory calculations .
annealed at temperatures around @xmath0c , fe@xmath1ni@xmath2 films on w(110 ) break into micron - sized bcc and fcc domains with compositions in agreement with the bulk fe - ni phase diagram .
ni is found to be the diffusing species in forming the chemical heterogeneity .
the experimentally - determined energy barrier of 1.59 @xmath3 0.09 ev is identified as the vacancy formation energy via density functional theory calculations .
thus , the principal role of the surface in the phase separation process is attributed to vacancy creation without interstitials . |
brans - dicke theory@xcite remained amongst the most talked about relativistic theory of gravity after general relativity .
soon after the theory was brought into being , a conformal transformation was suggested by dicke@xcite which can recast the theory in a different frame , called einstein frame where the field equations look more tractable .
the original theory , given in a frame , popularly dubbed as the jordan frame , is a manifestly non - minimally coupled theory where a long range scalar field @xmath0 has an interference term with the curvature @xmath1 .
the term looks like @xmath2 in the action
. a suitable conformal transformation of the form @xmath3 breaks this non - minimal coupling so that the scalar field contributes only through the kinetic term in the action .
the matter part also now depends on the scalar field as the stress energy tensor goes through a corresponding transformation@xcite .
this dependence does not formally affect the calculations unless one goes on to solve the geodesic equation .
this conformally transformed frame is called the einstein frame . brans - dicke ( bd )
theory indicates that the newtonian constant @xmath4 effectively is a function of the scalar field @xmath0 as @xmath5 , where @xmath6 is a constant and can be taken to be the present value of @xmath4 in a cosmological context . in the so called einstein frame
, @xmath4 regains its status of being a universal constant but the price one has to pay is that the rest mass of the test particle becomes a function of the scalar field @xmath0 and hence one has to forgo the principle of equivalence . for a compact review of this and some related issues
, we refer to the work of morganstern@xcite . + the intriguing question that is still alive is whether this two descriptions are equivalent or not .
cho@xcite argued that the jordan frame is not really the physical description of gravity and the same conclusion holds for kaluza - klein theory .
faraoni and gunzig@xcite arrived at the result that , within the realm of classical framework , einstein frame is more trustworthy than the jordan frame .
their work was based upon considerations of gravitational waves .
chiba and yamaguchi discussed the frame dependence of various cosmological parameters@xcite . in a more recent work , faraoni and nadeau@xcite , however , show that the with some proper interpretation , the two versions are actually equivalent at the classical level .
salgado@xcite also showed that the apparent mismatch of the cauchy problem in the two versions can actually be resolved . + it is quite expected that the two versions do not give the same physics as they are based upon different physical principles , the principle of equivalence holds in one and does not in the other .
notwithstanding the question of which version is better , practising relativists prefer to work in the jordan frame so as to be in the comfort zone of the principle of equivalence . for a computational advantage
, one might opt for the einstein frame but while discussing the physical aspects , the metric is transformed back to the jordan frame via the inverse transformation provided there is no singularity in the solution of the scalar field .
+ now the bigger question arises regarding the equivalence of the physics obtained in one frame and that in the same frame when the solutions are actually transformed back from the other .
the question might appear superfluous , but deserves attention in view of the high degree of nonlinearity in einstein systems . in the present work
, we deal with the question whether these two frames are equivalent even when they are looked at in the same jordan frame at the quantum level . in the two frames they would look different ,
but if we transform the solutions in the einstein frame back to the jordan frame by effecting the inverse transformation at the level of the solution , would they look the same ?
we find an answer in the negative !
we quantize a spatially flat friedmann - robertson - walker ( frw ) cosmological model with a perfect fluid in brans - dicke theory in both the versions and pretend that these two are different models altogether .
then we transform the solution for the wave - packet in the einstein frame to the jordan frame via the inverse conformal transformation and check if these two results match .
it is quite clearly observed that the results are different .
it would have been nice to work only with a bd field , i.e. , without a fluid content , but the evolution of the fluid provides a meaningful choice of a properly oriented time parameter , so the evolution of the system obtained is indeed physically relevant .
+ recently there has been a similar result , given by artymowski , ma and zhang@xcite , in loop quantum cosmology .
it is shown that the quantized version of a spatially flat frw model in bd theory either in vacuum or in the case of an additional scalar field indeed shows different behaviour in the two frames in the formalism of loop quantum cosmology .
+ in the following section we present the quantum bd model in the jordan frame and in section 3 the same in einstein frame . in the last section
we discuss the conclusions drawn from the present work .
the relevant action for brans - dicke theory with a perfect fluid can be written as @xmath7 + \int d^4x \sqrt{-g } ~p~~,\ ] ] where @xmath1 is the ricci scalar , and @xmath0 is the bd scalar field which is non - minimally coupled to gravity and @xmath8 is the dimensionless bd parameter . here
units are so chosen that @xmath9 .
the last term in the equation ( 1 ) represents the perfect fluid contribution to the action where @xmath10 is the pressure and is related to the energy density by the equation of state @xmath11 where @xmath12 . this restriction on @xmath13 stems from the consideration that sound waves can not propagate faster than light .
we shall work in a spatially flat spacetime given by the metric @xmath14 where n(t ) is called the lapse function , and @xmath15 is the flat 3-space metric . in schutz
s formalism @xcite , the fluid four velocity can be expressed in terms of some thermodynamic potentials . using the normalization of the velocity vector @xmath16
one can write the pressure @xmath10 in any spacetime without rotation as @xmath17 and the fluid part of the action takes the form ( in a comoving system where @xmath18 ) @xmath19 \ , .\ ] ] as none of the quantities mentioned depende on space coordinates , the spatial part of the volume integral @xmath20 yields a constant @xmath21 which is taken to be unity .
this does not infringe upon the generality as the constant will not contribute to the subsequaent variation . here
@xmath22 , @xmath23 , @xmath24 and @xmath25 are thermodynamic potentials which determine the velocity vector in schutz formalism . they all satisfy their own evolution equations . an overhead dot represents a differentiation with respect to the coordinate time @xmath26 .
the detailed method can be found in the work of lapchinskii and rubakov@xcite .
the method has been subsequently used with a high degree of usefulness by many , such as alvarenga _ et al_@xcite , vakili@xcite , majumder and banerjee@xcite , pal and banerjee@xcite . particularly , for a scalar field
, the method had been utilized by vakili@xcite , majumder@xcite and almeida _
. a very similar approach of expressing the fluid lagrangian in terms thermodynamic variables has been utilized very recently by b@xmath27hmer , tamanini and wright@xcite and koivisto , saridakis and tamanini@xcite .
+ we effect the canonical transformations , @xmath28 along with @xmath29 . here
@xmath30 , @xmath31 and @xmath32 , the lagrangian density of the fluid , is the expression inside the square bracket of equation ( [ af ] ) .
the corresponding hamiltonian for this perfect fluid can be written as @xmath33 the advantage of using this canonically transformed version is that we could find a set of variables where the system of equations is a lot more tractable , while the hamiltonian structure of the system remains intact . for a discussion regarding the hamiltonian structure in terms of poisson brackets ,
see ref @xcite . for the spatially flat frw model given by the metric ( [ frw ] )
, the ricci scalar can be written as @xmath34 ~~,\ ] ] where an overhead dot indicates a differentiation with respect to time @xmath26 . with this @xmath1 ,
the lagrangian density @xmath35 for the gravity sector becomes @xmath36 using a pair of new variables @xmath37 and @xmath38 in place of @xmath39 and @xmath0 , given by , @xmath40 and @xmath41 where @xmath42 and @xmath43 are constants , one can write @xmath35 as @xmath44 ~~.\end{aligned}\ ] ] it deserves mention that this transformation of variables from ( @xmath45 ) to ( @xmath46 ) is not a canonical transformation and one has to transform back to the old variables for any physical interpretation .
this transformation is effected only to facilitate a separation of variables .
+ in what follows , we shall work with a particular choice @xmath47 for the bd parameter @xmath8 .
there is no significance of this value , this is done only to facilitate the integration so that we can talk about an analytical solution of the wheeler - dewitt equation .
the hamiltonian for the gravity sector can be found out from the expression for the lagrangian @xmath35 and the net or super hamiltonian for the minisuperspace can be written as @xmath48 ~~.\end{aligned}\ ] ] here @xmath49 acts as a lagrange multiplier taking care of the classical constraint equation @xmath50 . using the usual quantization procedure @xcite
, we write the wheeler - dewitt equation for our super hamiltonian with the ansatz that the super hamiltonian operator annihilates the wave function , @xmath51 we now promote the variables to operators such as @xmath52 , @xmath53 with a particular choice of operator ordering we solve the eqn ( [ wdwe1 ] ) by the method of separation of variables .
it is analytically easy to solve the wheeler - dewitt equation with @xmath54 and @xmath55 .
* @xmath54 , @xmath56 and @xmath57 @xmath58 \nonumber \\ \left [ c_3 j_{\frac{k}{c}}\left ( \sqrt{\frac{16e}{3 } } e^{cq } \right ) + c_4 j_{-\frac{k}{c}}\left ( \sqrt{\frac{16e}{3 } } e^{cq } \right ) \right]~~. \end{aligned}\ ] ] here @xmath59 and @xmath60 are the constants from the separation of variables and @xmath61 are integration constants with @xmath62 as the bessel function of first kind .
if written as a function of @xmath39 , @xmath0 and @xmath63 , we get + @xmath64 \times \nonumber \\
\left [ c_3 j_{\frac{k}{6b}}\left ( \sqrt{\frac{16e}{3 } } a^3\sqrt{\phi } \right ) + c_4 j_{-\frac{k}{6b}}\left ( \sqrt{\frac{16e}{3 } } a^3\sqrt{\phi } \right ) \right].\end{aligned}\ ] ] * @xmath55 , @xmath65 and @xmath66 @xmath67 \nonumber \\ \left [ c_3 j_{\frac{k}{c}}\left ( \sqrt{\frac{40e}{3 } } e^{cq } \right ) + c_4 j_{-\frac{k}{c}}\left ( \sqrt{\frac{40e}{3 } } e^{cq } \right ) \right ] .
\end{aligned}\ ] ] as a function of @xmath39 , @xmath0 and @xmath63 , this becomes + @xmath68 \times \nonumber \\ \left [ c_3 j_{\frac{k}{3b}}\left ( \sqrt{\frac{40e}{3 } } a^{\frac{3}{2}}\sqrt{\phi } \right ) + c_4 j_{-\frac{k}{3b}}\left ( \sqrt{\frac{40e}{3 } } a^{\frac{3}{2}}\sqrt{\phi } \right ) \right ] .
\label{nordust}\end{aligned}\ ] ] the solution for the wave - packet can be obtained by a linear superposition of the eigenfunctions .
we do that for a dust distribution ( @xmath69 ) for the second case ( @xmath65 , @xmath66 ) and get , @xmath70 here we have considered @xmath71 ( from eqn ( [ nordust ] ) ) to satisfy the boundary condition imposed on @xmath72 . a change in variable @xmath73 is also considered .
the function @xmath74 is a suitable weight function for the construction of the wave packet . if we consider @xmath75 , where @xmath76 is a positive quantity .
the integrals of ( [ wp1 ] ) can be analytically evaluated to yield @xcite .
@xmath77 first we integrate with respect to @xmath78 using the known integral for the bessel function @xmath79 such that the second integral becomes a known gaussian integral . here
an overall constant accompanying @xmath81 , coming from several steps in the calculation , is taken to be unity without any loss of generality .
we can also calculate @xmath82 , the norm of the wave packet as @xmath83 where @xmath84 @xmath85^{\frac{1}{2 } } ~~ , \nonumber\ ] ] and @xmath86^{\frac{1}{2 } } ~~. \end{aligned}\ ] ] in fig.([modpsi ] ) we plot the nature of @xmath82 as a function of @xmath39 and @xmath0 .
the relevant action for the theory with a perfect fluid can be written as @xmath87 \nonumber \\ + \int d^4x \sqrt{-\bar{g } } ~{\bar p}~~,\end{aligned}\ ] ] where @xmath88 is the ricci scalar , and @xmath89 is a massless scalar field which is minimally coupled to gravity in the revised version of the theory .
one can arrive at this action with the aid of a conformal transformation @xmath90 the scalar field @xmath91 the quantities appearing in this action are in the transformed version , an overhead bar indicates that .
+ in a flat frw background it is rather easy to calculate the super hamiltonian constraint and it can be written as @xmath92 the coefficient of @xmath93 contains an additional @xmath94 as we shall work with the coordinates @xmath95 of the jordan frame so that the final comparison is consistent .
this hamiltonian is consistent with that given by zonghong@xcite ( see also @xcite ) . for the relevant transformations ,
we refer to @xcite . here
also the schutz formalism has been utilized .
following the same procedure as before , we solve the wheeler - de witt equation with @xmath54 and @xmath55 .
the stationary wave functions in the revised version are given as : * @xmath54 and @xmath57 @xmath96\times \nonumber \\ \left [ c_3 j_{\frac{\sqrt{1 + 3k^2}}{6}}\left ( \sqrt{\frac{8e}{3 } } \bar{a}^3 \right ) + c_4 j_{\frac{-\sqrt{1 + 3k^2}}{6}}\left ( \sqrt{\frac{8e}{3 } } \bar{a}^3 \right ) \right]\end{aligned}\ ] ] * @xmath55 and @xmath66 @xmath96 \times \nonumber \\ \left [ c_3 j_{\frac{\sqrt{5 + 24k^2}}{3\sqrt{5}}}\left ( \sqrt{\frac{32e}{3 } } \bar{a}^{\frac{3}{2 } } \right ) + c_4 j_{-\frac{\sqrt{5 + 24k^2}}{3\sqrt{5}}}\left ( \sqrt{\frac{32e}{3 } } \bar{a}^{\frac{3}{2 } } \right ) \right]\end{aligned}\ ] ] we have used some of the notations ( @xmath97 ) same as that of the last section , but they are not same and only related to the respective differential equation .
we take up the @xmath55 case in this transformed version as well .
we will consider @xmath98 so as to have the similar boundary conditions like the example in the jordan frame and get @xmath99 as @xmath100 and @xmath101 .
the wave - packet can be constructed by the superposition of the eigenfunctions as @xmath102 where @xmath103 , @xmath104 and we have incorporated a quasi - gaussian weight factor . if we make the approximation @xmath105 and @xmath106 then the above integral can be evaluated in a closed form @xcite . the approximations are only meant to evaluate the integrals analytically such that we can compare with results found in the last section . upto a constant proportionality factor , which comes from several steps of integration , we get @xmath107 where , using the inverse transformation , @xmath108 and @xmath89 are replaced by @xmath39 and @xmath0 respectively .
we can also calculate @xmath109 aimed for finding the norm of the wave packet , @xmath110 where @xmath111 @xmath112^{\frac{1}{2 } } ~~ , \nonumber\ ] ] and @xmath113^{\frac{1}{2 } } ~~.\ ] ] in fig.([modpsi2 ] ) we plot the nature of @xmath82 as a function of @xmath39 and @xmath0 .
if we compare equations ( 20 ) and ( 31 ) , the solution for the wheeler dewitt equations for the dust distribution ( @xmath69 ) for a particular value of the bd parameter @xmath8 as @xmath114 , we find that the solutions are evidently different .
one should emphasize that this comparison is made after the einstein frame solution is transformed back to the jordan frame via the inverse transformation , @xmath115 .
the solutions are intricate , so it is perhaps better to look at some physical aspects of the solutions .
the norm of the wave packets are found in the two versions .
they are given by equations ( 23 ) and ( 34 ) for the one that calculated in the jordan frame directly and the one that calculated in the jordan frame from the solutions transformed back from the einstein frame respectively .
the norms are depicted in figures ( [ modpsi ] ) and ( [ modpsi2 ] ) respectively for two epochs of time .
it is once again quite evident from the figures that the corresponding norms are qualitatively different , peaked at different locations .
thus it is conclusively established that the two frames are physically different at least at the quantum level .
it is true that this work shows this in one example .
but one counter example is good enough to show the non - equivalence of the two frames .
the particular values of @xmath13 and @xmath8 chosen are indeed for the sake of analytical calculations , but the values are quite legitimate , @xmath69 represents a pressureless dust , and @xmath66 is neither pathological nor trivial .
in fact , as evident from the action functional ( 27 ) , @xmath116 is the trivial case as it indicates a zero kinetic energy for the bd field and values of @xmath8 less than that would indicate a pathology of a negative kinetic energy . we have solved the wheeler dewitt equation for another case as well ,
namely @xmath117 indicating an effective cosmological constant with another legitimate value of @xmath118 but the comparison of the two frames were not possible for the difficulty in the integration .
however , if we compare the solution of the wheeler - dewitt equations , namely the solution given in equation ( 18 ) with that in ( 29 ) , it is rather apparent that they would not match after effecting the inverse transformation in the latter .
this indeed lends a support towards the claim of the present work regarding the nonequivalence of the two frames . in agreement with the recent result in loop quantum cosmology@xcite
, the present work , in a different philosophy of quantization , explicitly shows that jordan frame and einstein frames are physically different .
our result is also consistent with the finding of faraoni and nadeau@xcite that the two frames are not equivalent at the quantum level .
certainly there is ample scope of improvement on the present work .
an important issue is that there is nothing to check that the hamiltonian written in the two frames correspond to each other so far as the operator ordering is concerned .
the only thing that could be taken care of in terms of correspondence is the fact that in both cases the functions of the coordinates come first followed by the momenta .
we also took care to fix the boundary conditions in some way .
but question remains if that is enough .
+ but the question of this equivalence is important , and until this is settled , this kind of indications have to be relied upon in the absence of better results . + *
acknowledgement : * the authors are grateful to ananimous referee for many important and thought provoking comments which helped us in improving the manuscript by a great deal .
+ 100 c. brans and r.h .
dicke , phys .
, * 124 * , ( 1961 ) 925 .
dicke , phys .
, * 125 * , ( 1962 ) 2163 .
morganstern , phys .
d , * 4 * , ( 1970 ) 946 .
cho , phys .
* 68 * , ( 1992 ) 3133 .
v. faraoni and e. gunzig , arxiv : astro - ph/9910176 . t. chiba and m. yamaguchi , jcap10(2013)040 .
v. faraoni and s. nadeau , phys .
d , * 75 * , ( 2007 ) , 023501 .
m. salgado , class .
* 23 * , ( 2006 ) , 4791 .
m. artymowski , y. ma and x. zhang , phys .
d , * 88 * , ( 2013 ) 104010 .
schutz , phys .
d2 * , ( 1970 ) 2762 .
schutz , phys .
* d4 * , ( 1971 ) 3559 .
lapchinskii and v.a .
rubakov , theor . math . phys .
* 33 * , ( 1977 ) 1076 .
f.g.alvarenga , j.c.fabris , n.a.lemos and g.a.monerat , gen .
rel . grav . * 34 * , ( 2002 ) 651 .
alvarenga , a.b .
batista , j.c .
fabris and s.v.b .
goncalves , gen .
35 * , ( 2003 ) 1659 .
b. vakili , class .
* 27 * , 025008 ( 2010 ) .
b. majumder and n. banerjee , gen .
. grav . * 45 * , ( 2013 ) 1 .
s. pal and n. banerjee , phys . rev .
d , * 90 * , ( 2014 ) 104001 .
s. pal an n.banerjee , arxiv : 1411.1167 [ gr - qc ] .
b. vakili , phys .
b , * 688 * , ( 2010 ) , 129 . b. majumder , phys . lett .
b , * 697 * , ( 2011 ) 101 . c.r .
almeida , b. batista , j.c .
fabris and r.l.v .
moniz , arxiv:1501.04170 [ gr - qc ] .
b@xmath27hmer , n. tamanini and m. wright , phys .
d , * 91 * , ( 2015 ) , 123002 .
koivisto , e.n .
saridakis and n. tamanini , arxiv:1505.07556 [ astro-ph.co ] charles w. misner , phys .
* 186 * , ( 1969 ) 1319 .
bryce s. dewitt , phys . rev .
* 160 * , ( 1967 ) 1113 .
m. abramowitz and i. a. stegun , _ handbook of mathematical functions _
, new york : dover ( 1972 ) .
z. zonghong , chinese phys .
lett . , * 9 * , ( 1992 ) , 273 . z. zonghong , h. chaoguang and l. liao , chinese phys .
lett . , * * , ( 1990 ) , 477 . | with an explicit example , we show that jordan frame and the conformally transformed einstein frames clearly lead to different physics for a non - minimally coupled theory of gravity , namely brans - dicke theory , at least at the quantum level .
the example taken up is the spatially flat friedmann cosmology in brans - dicke theory
. einstein frame , jordan frame , quantum cosmology , schutz formalism |
the study of globular cluster systems of dwarf galaxies complements the numerous studies of such systems in larger elliptical and spiral galaxies .
few globular cluster systems ( gcss ) around dwarf galaxies beyond the local group have been studied to date with respect to their cluster system ( see ashman & zepf 1998 ) .
this is mostly due to the low numbers of globular clusters present in such galaxies .
however , their properties are relevant for a number of globular cluster system formation scenarios .
dwarf galaxies are expected to provide insight into how the smallest galaxies build up a system of globular clusters .
further , their properties must be known in order to verify scenarios in which larger globular cluster systems are predicted to build up by the accretion of proto - galactic fragments or dwarf galaxies ( @xcite 1998a , @xcite 1998 , @xcite 1999 ) .
these scenarios relate to the older idea that galaxy halos might have formed through the assembly of such small stellar systems ( e.g. * ? ? ?
photometric studies of several globular cluster systems in dwarf galaxies were carried out by @xcite and @xcite .
@xcite studied the systems of 11 dwarf galaxies in the virgo cluster .
all were found to host globular cluster candidates and have specific frequencies ranging from 3 to 8 , similar to local group dwarfs and giant elliptical galaxies .
@xcite studied 24 dwarf ellipticals in the virgo and fornax clusters as well as in the leo group .
they found that de , n galaxies had higher specific frequencies than de galaxies , with values around @xmath9 , increasing with increasing @xmath10 ( decreasing luminosity ) . not much is known yet about the metallicities of globular clusters in dwarf galaxies .
@xcite constructed a metallicity distribution for all local group dwarf galaxies and noticed that the distribution was peaked around [ fe / h]@xmath11 dex with no clusters more metal - rich than [ fe / h]@xmath12 dex .
@xcite derived metallicities from washington colors for two of their virgo dwarf ellipticals and obtained a mean metallicity of [ fe / h]@xmath13 dex .
@xcite studied the gcs of the de ngc 3115 dw1 in more detail and found it to be relatively rich ( see below for a more detailed description of their results ) .
this motivated us to carry out spectroscopy for some of the globular cluster candidates in this galaxy to get a more detailed picture of their chemical and kinematical properties .
further , hst / wfpc2 data were available from the archive , allowing us to study the sizes of some of the clusters .
ngc 3115 dw1 is a de1,n galaxy in the vicinity of the giant s0 galaxy ngc 3115 .
it is located ra : 10h 05 m 41.6s ; dec : @xmath14 58 53.5 ( @xmath15 ; @xmath16 ) .
we will assume a distance of @xmath17 mpc throughout this paper following @xcite .
additional properties will be given in the text where they are relevant .
in section 2 we describe the new data . in section 3
we analyze the spectroscopic data giving first a brief kinematical study of the globular cluster system before discussing the abundances and the overall metallicity of the system . in section 4
we revisit the previous photometry ( number of clusters , colors , specific frequency ) , compare the photometric metallicities with the spectroscopic ones and add the hst / wfpc2 imaging to derive sizes for two clusters . in section 5
we discuss whether ngc 3115 dw1 could have suffered stripping by its giant companion .
we summarize our results in section 6 .
46 spectra of candidate globular clusters in ngc 3115 dw1 were obtained with the lris - a spectrograph @xcite on the keck - ii telescope during the nights of 1996 december 18th and 19th .
the observations were performed in multi - slit mode using two slit masks containing 24 ( mask a ) and 22 slits ( mask b ) , respectively .
mask a was exposed for 4800 seconds while mask b was exposed 6600 seconds .
the multi - object - spectra images were taken with a tektronix 2048@xmath182048 pix@xmath19 chip .
the seeing was around 0.8 during both nights .
for all observations the first order of the 600/5000 grism ( i.e. 600 lines mm@xmath20 ) was used with slitlets of 1 which provided a dispersion of 1.3 pix@xmath20 ( at 5000 ) and a spectral resolution of 4.1 fwhm@xmath20 .
all frames were binned ( 1@xmath21 ) perpendicular to the dispersion axis during the read - out . for each night
the images were reduced individually and eventually combined .
a master bias was created from 5 zero - second images , taken at the end of each night .
a master flat was obtained from 5 twilight - flat images obtained each night .
each object image was bias subtracted and flat - fielded in a standard way . with the iraf package _ apall _ we traced , extracted and sky subtracted all object spectra .
the sky was `` optimally '' subtracted , i.e. modeled with variance weighting perpendicular to the object spectra before being subtracted @xcite .
hgkrne - calibration - lamp spectra were obtained for wavelength calibration .
the calibration spectra were traced , extracted , and sky subtracted exactly in the same way as the object spectra .
the wavelength calibration was verified on sky spectra included in each slit spectrum .
the overall standard deviation of the wavelength calibration was determined to be @xmath22 for both nights .
the field - of - view of lris - a is 6 @xmath18 8 .
we placed the multi - slit masks on the central / north and south - eastern part of ngc 3115 dw1 as can be seen in fig .
[ ps : fov ] .
the slit masks were aligned such that the slits were pointing towards the center of the giant s0 galaxy ngc 3115 . due to the restrictions in making the masks ( long enough slits , wavelength coverage , etc . ) about 50% of the targets could be selected from the previous photometry of @xcite ( see fig .
[ ps : fov ] ) .
the other 50% was ` blindly ' selected from an lris acquisition image .
only 6 of the former group lay within 48 of the galaxy center , i.e. within the region where @xcite detected a clear overabundance of objects .
we therefore expected @xmath23 6 + objects in our total sample to be true globular clusters . in order to measure sizes of globular cluster candidates that were confirmed by our spectroscopy
, additional data for the globular cluster system were obtained from the hst archive .
short exposures of 160 seconds in f555w and 320 seconds in f814w filter were taken from program go:5999 ( pi : a.phillips ) .
a simple reduction procedure was applied to all hst images : the images were biased , flat - fielded and calibrated as described in @xcite , including corrections accounting for cte and field distortions .
the brightest clusters were matched on both the f555w and f814w images and their sizes were studied .
in addition , we re - analyzed the ground - based b , v data from @xcite .
the field - of - views of both photometric datasets are over - plotted in fig .
[ ps : fov ] .
in this section we extract heliocentric radial velocities @xmath24 of globular clusters and perform a mass estimate of the host galaxy , ngc 3115 dw1 .
subsequently , we measure abundances using the lick / ids passband definitions and infer a mean metallicity of the gcs .
we derived radial velocities by two different methods .
velocities were derived by cross - correlation with high s / n template spectra of two bright gcs in m31 : 225 - 280 and 158 - 213 ( for nomenclature see * ? ? ? * ) .
the cross - correlation was carried out with the iraf task _ fxcor_. all measurements are summarized in table [ tab : rvs ] .
the `` internal '' errors which are given by the cross - correlation code lie about @xmath25 km s@xmath20 for well defined high - s / n spectra and increase rapidly as the object spectrum becomes less defined . for 7 high - s
/ n spectra we also estimated the radial velocity by measuring redshifts of individual absorption lines ( see footnote in table [ tab : rvs ] ) .
the mean radial velocity from cross - correlation served as a first guess to the mean wavelength - shift determination .
we used the iraf package _ rvidlines _ which employs a _
center1d _ code to match the center of each individual absorption feature ( see manual of _ rvidlines _ ) .
we obtained an `` internal '' mean error from the averaging process which is @xmath23 100 km s@xmath20 . the results and their errors
are included in table [ tab : rvs ] . after obtaining the radial velocities from cross - correlation and in some cases from wavelength shifts of individual absorption lines , we combined all available measurements for further selection of globular clusters by radial velocity .
velocities derived from cross - correlation were given twice the weight of velocities from absorption lines .
the average radial velocity for each object can be found in the last but one column of table [ tab : rvs ] .
figure [ ps : rvhisto ] shows a radial - velocity histogram of objects with @xmath26 km s@xmath20 .
the figure shows a hint of bimodality due to the extended tail of velocities @xmath27 500 km s@xmath20 .
based on the measured radial velocity of ngc 3115 dw1 of @xmath28 km s@xmath20 ( @xmath29 km s@xmath20 @xcite 1991 , @xmath30 km s@xmath20 @xcite 1993 , and @xmath31 km s@xmath20 @xcite 1993 ) and an adopted velocity dispersion of the gcs in ngc 3115 dw1 of @xmath32 km s@xmath20 ( see below ) , globular clusters should have radial velocities in the range @xmath33 km s@xmath20 .
seven candidates lie within this range .
this cut enables a reliable differentiation between foreground objects of low radial velocity and globular clusters in ngc 3115 dw1 .
we used the kmm code @xcite to obtain the statistical significance of bimodality .
the kmm code fits two gaussians to the data using maximum - likelihood techniques .
the data can be fit with two gaussians of identical dispersion or two gaussians of independent dispersion .
we applied both techniques .
the mean radial velocities of the two gaussians are @xmath34 km s@xmath20 and @xmath35 km s@xmath20 for identical - dispersion and @xmath36 km s@xmath20 and @xmath37 km s@xmath20 for independent - dispersion fitting .
the error is the error of the mean calculated from the variance of each distribution .
kmm estimates also the fraction of data points which are part of each sub - distribution . of 28 objects in the histogram ( one additional object , d26 ,
is not included due to @xmath38 km s@xmath20 ) in figure [ ps : rvhisto ] , kmm assigns 21 objects to the sub - population with the lower mean radial velocity and 7 to the sub - population with the higher mean radial velocity which is in good agreement with the expectations ( see [ ln : specobs ] ) .
the confidence level for bimodality is @xmath39% .
the ( weighted ) mean radial velocity of all 7 detected globular clusters is @xmath40 km s@xmath20 .
this value deviates from the measured radial velocity of ngc 3115 dw by @xmath41 which is a hint that we are biased towards lower velocities by both the small sample and the choice of slit - mask position on the sky .
since we detected globular clusters predominantly in the central and northern field of ngc 3115 dw1 our measurements could be influenced by a systematic rotation of the gcs ( see also sect . [
ln : rotdisp ] ) .
all 22 objects with low radial velocities have a weighted mean velocity of @xmath42 km s@xmath20 . assuming a simple stellar rotation model for the milky way we expect the foreground stars in the direction of ngc 3115 dw1 ( @xmath15 and @xmath16 ) to have @xmath43 km s@xmath20 @xcite which is in rough agreement with our measurement .
in summary , within our data of 46 spectroscopically - analyzed objects we found 7 globular clusters 22 foreground stars , and 15 background galaxies ( 9 of them are significantly clumped about @xmath44 km s@xmath20 or @xmath45 ; the velocity dispersion of this potential galaxy cluster is @xmath46 km s@xmath20 ) .
2 objects ( d21,l6 ) could not be identified reliably and were therefore dropped .
the radial - velocity data for all objects are summarized in table [ tab : rvs ] .
our data sample of 7 globular clusters is only sufficient for a first rough mass estimate of ngc 3115 dw1 and its m / l ratio .
furthermore , we rely in this section on the assumption that the system is not influenced by the nearby giant s0 galaxy , an assumption that we will question in sec .
[ ln : discussion ] .
we used two mass estimators which are extensively described and tested by @xcite and @xcite .
we applied the virial mass estimator ( vme ) and the projected mass estimator ( pme ) accounting for different orbit characteristics of the globular clusters .
since the alignment of our multi - slit masks is biased towards clusters of the central and northern quadrant of the galaxy , our data is subject to unknown systematic east - west rotation of the entire globular cluster system . if the angular momentum vector points along the north - south axis we will have under- or overestimated the mass depending on the direction of rotation . although , it is possible to provide a lower - mass limit from our spatially constrained cluster sample by varying the mean systemic velocity ( see below ) , it is not possible to correct completely for the unknown total rotation of the globular cluster system ( see also [ ln : rotdisp ] ) .
we let the radial velocity of ngc 3115 dw1 ( which is assumed to be the mean velocity of the gcs as well ) vary over a wide range while the measured radial velocities of the globular clusters remained fixed . during each step of 10 km s@xmath20 we calculated the mass of the galaxy with each mass estimator .
the resulting plot ( mass vs. @xmath24 of the gcs ) is shown in figure [ ps : massestim ] .
table [ tab : massestimate ] shows the lower mass limits calculated with all mass estimators . at the highest measured radial velocity of ngc 3115 dw1 ( @xmath47 km s@xmath20 , * ? ? ?
* ) we obtain total masses in the range @xmath48m@xmath3 to @xmath49m@xmath3 .
the estimated masses for the measured mean radial velocity of the globular cluster system ( @xmath50 km s@xmath20 ) are in the range @xmath51m@xmath3 to @xmath52m@xmath3 .
the errors in table [ tab : massestimate ] are the intrinsic , statistic , and systematic uncertainties of the mass - estimate .
the first is the uncertainty of the code itself @xcite while the remaining are due to the limited sample size and the uncertain distance .
all masses are the total mass estimates within a galactocentric radius of @xmath53 or @xmath54 kpc , respectively .
this is the projected radial distance of the outermost globular cluster ( l63 ) .
assuming an _ isotropic _ orbit distribution for globular clusters in ngc 3115 dw1 the lower - mass limit for the galaxy of @xmath55m@xmath3 seems rather large for a dwarf elliptical .
the absolute magnitude of @xmath8 mag @xcite is high as well ( more than a magnitude brighter than m32 for example , and similar to ngc 4486b , although ngc 3115 dw1 has a dissimilar structure to that of these low - l es ) .
the mean absolute v - magnitude for nearby dwarf ellipticals is @xmath56 mag @xcite .
considering its mass and luminosity we address the fact that ngc 3115 dw1 appears to be a transition - type galaxy between luminous des and low - luminosity ellipticals in the discussion section . limiting the data set of radial velocities to smaller radii we can probe the radial mass dependencies in ngc 3115 dw1 . clearly , statistical errors become important when we reduce the already small data set by removing the outermost globular clusters .
nonetheless , we calculate mass estimates for different radii ( using the pme and assuming isotropic orbits ) since the kinematics of the two outer clusters might be influenced by the nearby giant s0 .
the results are summarized in table [ tab : massrad ] . combining the former findings with the photometry of @xcite , who found total magnitudes @xmath57 mag and @xmath58 mag , we estimate rough mass - to - light ratios for different radii . applying a king profile @xcite with the parameters
@xmath5914.4 and @xmath601.4 @xcite we obtain at the outermost globular - cluster projected radius @xmath61 ( @xmath62 kpc ) m / l@xmath63 with a @xmath64 uncertainty due to statistical sample - size uncertainties and photometric errors . the systematic error due to distance uncertainties of ngc 3115 dw1 is @xmath65 . going inwards , the m / l@xmath66 ratio drops . at a projected radius of @xmath67 ( @xmath68 kpc )
the mass - to - light ratio is m / l@xmath69 .
our analysis can be expanded by combining our results with the m / l@xmath66 measurements for the innermost part of ngc 3115 dw1 .
@xcite measure within @xmath70 ( or @xmath71 pc ) of ngc 3115 dw1 a central velocity dispersion of @xmath72 km s@xmath20 .
they derive a m / l@xmath66 of @xmath73 within @xmath70 . a gradient of m / l@xmath66
can therefore be traced outward from the center of the galaxy , although the uncertainties are quite large .
however , all values fit well into the range given by @xcite for dwarf elliptical galaxies ( m / l@xmath74 for fornax up to m / l@xmath75 for ursa minor ) . using the maximum - likelihood method of @xcite we measure a marginal net rotation of @xmath76 km s@xmath20 for the globular cluster system of ngc 3115 dw1 .
note that to date no de was reported to show significant rotation of its stellar body @xcite .
the position angle of the rotation axis is @xmath77 ( poorly defined given the weak rotation ) .
this result may be of course biased by the incomplete spatial coverage of our small sample .
correcting for the net rotation we obtain a line - of - sight velocity dispersion of @xmath78 km s@xmath20 for the full sample of 7 clusters , and @xmath79 km s@xmath20 for the inner 5 clusters .
the major axis of the de1,n galaxy of @xmath80 @xcite appears at face value nearly _
parallel _ to the rotation axis of the gcs although the latter is not well defined as mentioned above .
spectroscopy for more clusters with better spatial coverage is certainly needed to establish the axis alignment .
the s / n of our spectra are insufficient to reliably establish individual cluster metallicities . to measure a mean abundance for
the gcs we combined all the individual spectra into a high - s / n ` mean ' spectrum .
all individual globular cluster spectra as well as the combined spectrum are shown in figure [ ps : spectra ] . for abundance measurements we used the passband definitions of @xcite and
the new lick / ids passband definitions of @xcite .
all abundances ( both of single spectra and the mean spectrum ) are given in table [ tab : abundmag ] .
@xcite calibrate single element abundances with [ fe / h ] using a large sample of milky way and m31 globular clusters .
we used their calibration to estimate the mean [ fe / h ] for the entire gcs , based on the composite spectrum .
the mean metallicity of our globular cluster sample is @xmath81[fe / h]@xmath82 dex .
exactly the same value is obtained from the weighted mean of the individual measurements ( see table [ tab : metalcomp ] ) .
overall , the mean metallicity does _ not _ follow the empirical gcs - metallicity galaxy - luminosity relation ( see fig . 5.7 in * ? ? ?
ngc 3115 dw1 ( @xmath8 mag ) falls in the transition region between dwarfs and elliptical galaxies , whereas the mean metallicity of the gcs falls in the range of metallicities found in giant elliptical galaxies .
according to this empirical relation ngc 3115 dw1 appears to be slightly too metal - rich for its luminosity .
we compare the mean abundance ratios of the gcs with abundance ratios of globular clusters in other galaxies .
@xcite provide a compilation of abundances of milky way and m31 globular clusters .
abundances of ngc 1399 globular clusters were measured by @xcite .
both data sets use the definitions of the lick / ids system . in order to minimize the statistical noise of abundance measurements
we calculate a mean iron abundance @xmath83 and a mean metal abundance @xmath84 $ ] ( see * ? ? ?
* for a detailed discussion ) .
figure [ ps : metals ] shows abundance ratios for several dominant elements .
the upper four panels show abundance ratios relative to the [ mgfe ] index . in the upper left panel
the age - sensitive h@xmath85 abundance is seen to be in good agreement with milky way and m31 data , although possibly at the lower ( older ) edge .
the upper right panel shows the g - band index compared with the [ mgfe ] index . the g - band is a primary metallicity indicator @xcite .
the data show no abundance anomalies .
the two middle panels of figure [ ps : metals ] could in principal be used to examine the @xmath86-element content of these globular clusters . for the stellar light of ( mostly brighter ) ellipticals @xcite found an @xmath86-element enhancement .
the [ @xmath86/fe ] enhancement is a very sensitive indicator of the star formation rate in a galaxy .
as @xmath86-elements are preferentially created in sne type ii , their enhancement indicates a violent star formation and/or a top - heavy imf .
a depression , or normal values , of [ @xmath86/fe ] would result from quiet star formation in which sne ia dominate the enrichment processes .
given the relatively high mean metallicity ( @xmath81[fe / h]@xmath82 dex ) , a normal @xmath86-element ratio would suggest that these clusters formed from enriched material either during epochs of quiet star formation , or at the very beginning of a burst .
better data could lead to interesting insights on this topic .
the main result from the photometry of @xcite was that the gcs in ngc 3115 dw1 is rich with a specific frequency of @xmath87 and a total globular cluster population @xmath88 .
our spectroscopy and the photometric data set of durrell et al .
have 22 objects in common ( see figure [ ps : fov ] and table [ tab : specphot ] ) .
6 of these 22 objects have projected radial distances of @xmath89 ( @xmath90 kpc ) .
durrell et al .
consider the gcs to lie within this radius ( mainly because the surface over - density of objects disappears beyond it ) .
we can confirm 4 of the 6 objects as bona - fide globular clusters ( d7 , d14 , d15 , and d46 ) .
one object was found to be a background galaxy and another can not be identified either by radial velocity or by its spectrum . assuming that this sample of 6 objects is a statistically representative sample of objects in the projected vicinity of ngc 3115 dw1 the upper limit of the probability of finding a globular cluster within a radius of @xmath89 ( @xmath90 kpc ) around the center of ngc 3115 dw1 is @xmath91 ( the lower limit is @xmath92 ; if we exclude the non - identified object ) .
this is in good agreement with the findings of @xcite , although the statistical significance is very low .
durrell et al .
measure the contaminating surface density of background objects to be @xmath93 arcmin@xmath94 . within their radial limit of @xmath89
there are a total of @xmath95 background objects . for a total population of @xmath96 , the probability of picking a globular cluster within @xmath89 ( and at the magnitude limit of the photometry of durrell et al . )
is @xmath97 .
our spectroscopic results suggest that there is no need to make any correction to the values for specific frequency and total globular cluster population size derived from photometry .
@xcite measure specific frequencies for dwarf elliptical galaxies in the virgo and fornax cluster and find a @xmath98 relations for the nucleated dwarfs .
ngc 3115 dw1 has a higher specific frequency ( @xmath99 ) than the @xmath100 derived from miller s et al .
relation for group and cluster de , n galaxies .
figure [ ps : cmd ] shows the cmd of objects with @xmath89 ( @xmath90 kpc ) which have been marked by open squares .
objects with photometry by durrell et al . , spectroscopically identified foreground stars , background galaxies , and globular clusters are indicated .
we determined the mean color of 5 globular clusters ( 2 spectroscopically confirmed globular clusters are not included in the photometric sample ) in this cmd with maximum - likelihood techniques .
we obtained @xmath101 mag with a dispersion of @xmath102 mag .
durrell et al .
found @xmath103 mag , @xmath104 mag for the total sample i.e. corresponding to a lower mean metallicity .
our subset seems to be slightly biased towards metal - rich objects . in order to constrain the significance of photometrically - derived metallicities
we transform the @xmath105 color into a [ fe / h]-metallicity and compare it with the findings of our abundance measurements . for this purpose
we use the relation of @xcite @xmath106}=5.0\cdot(b - v)_{\rm o}-4.86\ ] ] with [ fe / h ] being the independent parameter during the calibration of the equation .
the application of equation [ eq : colormetallin ] to all de - reddened ( e@xmath107 mag , * ? ? ?
* ) globular - cluster colors in our ngc 3115 dw1 sample leads to photometrically derived metallicities which can be compared with the [ fe / h ] values from spectroscopy .
all data are summarized in table [ tab : metalcomp ] .
the resulting weighted mean metallicity is @xmath81[fe / h]@xmath108 dex ( for the 5 globular clusters ) with a dispersion of @xmath109[fe / h]@xmath110 dex in good agreement with the values derived from spectroscopy .
the uncertainty results from the photometric error of the color only .
no transformation uncertainty was included .
we matched two spectroscopically - confirmed globular clusters , d15 and d25 ( see figure [ ps : fov ] and table [ tab : specphot ] ) , in the hst images taken from the archive ( both on wf chips ) .
at the distance of ngc 3115 dw1 the wf chips resolve globular clusters with core radii of @xmath1115 pc .
only 14% of milky - way globular clusters show core radii larger that 5 pc @xcite .
however , we can use the hst data to derive upper limits for the globular cluster sizes .
only images taken through the f814w filter were used because of their higher s / n .
the radial source profile @xmath112 ( i.e. the psf of the final image ) is a convolution of the object profile @xmath113 with the telescope psf @xmath114 and an additive noise term @xmath115 ( @xmath116 being the radial distance from the center of the profile ) ; @xmath117 to calculate the telescope - psf profile @xmath114 we used the tinytim v4.4 code by @xcite which gives a semi - analytic estimation of the hst - psf for each chip , each chip position , and each filter .
we adopted a king profile @xcite for @xmath113 which appears to be a good fit to globular - cluster radial profiles in milky way ( e.g. * ? ? ? * ) and extragalactic systems ( @xcite 1996 in m31 , @xcite 1985 in lmc , @xcite 1998 in ngc 3115 , @xcite 1999 in m87 , and @xcite 1999 in ngc 4472 ) . from an analysis of milky way globular clusters @xcite
we chose @xmath118 as a concentration parameter .
in equation [ eq : deconv ] we neglect the additive noise term since our size - estimation errors are dominated by the convolution of the poorly - defined charge - diffusion matrix with the optical hst - psf ( see * ? ? ? * for a detailed discussion ) .
note that the charge - diffusion smears 25% of the infalling light of the central pixel among its neighbors .
for consistency with other work we continue to use this convolution throughout our analysis despite the fact that the diffusion correction has been derived only for the f555w filter and is thought to be wavelength dependent .
five king profiles were generated with core radii in the range @xmath119 pix ( @xmath120 pc ) .
in addition , we generated hst - psfs for both our identified globular clusters using individual specifications ( e.g. filter , chip , chip position ) .
both hst - psfs were convolved with all the king profiles .
aperture photometry was applied to all generated profiles and both globular clusters on the hst images . for this purpose we used sextractor @xcite and
measured magnitudes in 30 apertures with diameters in the range 130 pix .
all magnitudes were normalized to the average aperture magnitudes in the range 1030 pix , i.e. @xmath121 .
figure [ ps : profile ] shows the profiles of both globular clusters , a raw hst - psf profile , and two of the convolution profiles with core radii of @xmath122 and 0.4 pix .
both globular cluster profiles deviate significantly from the raw hst - psf profile which indicates that both d15 and d25 are resolved .
the s / n of the f814w image drops to 1 at an aperture diameter of 6 pix . for larger apertures
there is insufficient signal to detect any deviations from a raw hst - psf . for smaller apertures
both globular cluster profiles lie between king profiles of core radii 0.2 pix and 0.4 pix .
we deduce an upper limit for both globular cluster sizes of @xmath123 pc at an adopted distance of ngc 3115 dw1 of @xmath124 mpc @xcite .
27% of milky - way globular clusters have core radii larger than 2.1 pc @xcite and 10% have sizes in the range defined by the errors of the ngc 3115 dw1 clusters .
this upper limit compares well with the results of @xcite and @xcite . using hst photometry
, these authors find typical half - light radii of @xmath125 pc and @xmath126 pc for globular clusters in ngc 3115 and m87 , respectively .
in section [ ln : rotdisp ] , we derived a high globular cluster velocity dispersion , and thus a high galaxy mass , when we included the two outermost globular clusters .
the high mass is not unexpected given the bright absolute magnitude of ngc 3115 dw1 . based on its @xmath127 mag @xcite , ngc 3115
dw1 falls in the transition region between dwarfs and ellipticals in the mass - luminosity relation of @xcite ( see fig . 3 therein ) .
its high mass ( @xmath128m@xmath3 ) and the high velocity dispersion ( @xmath78 km s@xmath20 , see below ) are more consistent with an elliptical galaxy .
we therefore discuss whether the two outermost clusters could be in the process of being stripped by the nearby giant s0 galaxy ngc 3115 .
figure [ ps : fov ] and table [ tab : specphot ] show that 2 ( l1 and l63 ) of the 7 globular clusters have significantly larger projected radii , i.e. 161.4 ( @xmath129 kpc , l1 ) and 189.4 ( @xmath130 kpc , l63 ) , than the `` inner '' ( @xmath131@xmath132 kpc ) globular clusters .
these large projected distances from ngc 3115 dw1 could be due to stripping by the nearby s0 galaxy ngc 3115 .
figure [ ps : localgroup ] shows the relative positions of ngc 3115 dw1 and ngc 3115 .
the projected distance between the two galaxies is 17.3 which corresponds to 55 kpc at the distance of @xmath13311 mpc .
the mean radial velocities of l1 and l63 are @xmath134 km s@xmath20 and @xmath135 km s@xmath20 , respectively .
only l1 shows a significant deviation from the systemic velocity of ngc 3115 dw1 ( @xmath136 km s@xmath20 , * ? ? ?
* ) and ngc 3115 ( @xmath137 km s@xmath20 , * ? ? ?
we expect no contamination from globular clusters of the nearby galaxy ngc 3115 .
@xcite found the surface density over - abundance of globular clusters around ngc 3115 ( power - law index of radial distribution @xmath138 ) disappearing at 6 radius from the center of ngc 3115 ( at a photometric limit of @xmath139 mag ) .
the globular clusters l1 and l63 have a radial distance to ngc 3115 of @xmath140 .
the projected gc surface density of the gcs of ngc 3115 at the position of these two clusters is @xmath141 arcmin@xmath94 .
the extrapolated gc surface density of ngc 3115 dw1 at this position lies between 0.2 and 6.9 gcs arcmin@xmath94 , given the large uncertainties on the density profile .
as the numbers are too small ( we only found 2 clusters to the north and 0 to the south ) it can not statistically be concluded whether the two globular clusters found in the northern field are chance detections or a statistically significant overabundance . assuming that both galaxies are roughly at the same distance
, we can estimate the dwarf galaxy s gravitational potential and the ratio of potentials of ngc 3115 dw1 and ngc 3115 .
both globular clusters are at about 1/5 of the distance separating ngc 3115 dw1 and ngc 3115 . as a rough estimate
, we assume that ngc 3115 dw1 and ngc 3115 have similar m / l@xmath66 . in this case
, the ratio of their @xmath10 s would imply that ngc 3115 has a mass 10 times larger than ngc 3115 dw1 .
hence , the gravitational potentials are comparable at the projected position of the distant globular clusters ( l1 and l63 ) . since the mass of ngc 3115 is likely to be higher than the adopted value ( assuming an extended dark matter halo )
the motion of both globular clusters is no longer dominated by the gravitational potential of ngc 3115 dw1 alone . both clusters
could then be considered as intergalactic globular clusters .
note that stripping of globular clusters appears to be common among interacting galaxies .
@xcite show that four globular clusters of the sagittarius dsph are in the process of being stripped by the milky way and are being added to its globular cluster system .
other studies have indicated that stripping may be important in galaxy clusters ( e.g. in the fornax cluster @xcite 1999 , @xcite 1999 ) .
however , there are no other ( optical ) hints of interaction from ngc 3115 dw1 s stellar light . @xcite found the isophotes to be consistent with little or no tidal disruption out to a projected radius of 60(corresponding to 3.2 kpc ) where their photometric errors start to dominate .
a simple test for the stripping hypothesis would be a wide - field study of the system in order to rule out ( spectroscopically ) the presence of any similar clusters around ngc 3115 dw1 . a look at the fundamental plane of dwarf elliptical galaxies ( e.g. * ? ? ?
* ) shows that ngc 3115 dw1 fits reasonably well into the relation for dwarf and giant elliptical galaxies , under their assumption of @xmath142 mag .
adopting the absolute magnitude of @xmath8 mag @xcite the galaxy falls slightly off the relation and would imply a higher velocity dispersion than measured in the central 3 . with the measured velocity dispersion of @xmath79 km s@xmath20 for the 5 globular clusters inside @xmath143 ( @xmath144 kpc ) we obtain from the fundamental - plane relation of @xcite an absolute magnitude of @xmath145 mag .
the measured velocity dispersion for the total sample of 7 clusters inside @xmath146 ( @xmath147 kpc ) of @xmath78 km s@xmath20 , would correspond to far brighter absolute magnitude ( @xmath148 mag ) than the measured @xmath8 mag .
this discrepancy can be explained by a close encounter and subsequent stripping of the dwarf galaxy s halo by the nearby s0 galaxy ngc 3115 .
stripping of outer halo regions might well have introduced violent perturbations and led to an enhanced velocity dispersion of the halo region ( which is traced by the globular clusters ) .
the relaxation time of such a system far exceeds the hubble time @xcite and therefore it is not possible to reject this scenario just from considerations of dynamical timescales .
alternatively , a high velocity dispersion in the outskirts of a galaxy could be due to a dark - matter dominated massive halo .
the outer parts of a number of lower - luminosity local group galaxies are known to be dominated by dark matter ( e.g. * ? ? ?
this picture could explain the fact that we measure an uncommonly high mass for a dwarf elliptical ( see sec .
[ ln : massestimate ] ) at a projected radius of 189.4 ( @xmath130 kpc ) .
we can not discriminate between the above possibilities at this point .
using lris multi - slit spectra we confirm , on the basis of their radial velocities , 7 of the 46 objects in our spectroscopic sample as bona - fide globular clusters associated with the bright ( @xmath8 mag ) de1,n galaxy , ngc 3115 dw1 .
we verify the findings of @xcite ( within a projected radius of @xmath89 corresponding to @xmath90 kpc ) who derived the specific frequency @xmath99 and a total globular cluster population size @xmath96 .
the spectroscopic verification of foreground and background contamination indicates that no revision of these results is necessary .
ngc 3115 dw1 remains a de , n galaxy with a somewhat high @xmath149 value .
a mass estimate using the projected mass estimator ( pme ) for isotropic globular cluster orbits yields a total galaxy mass of @xmath150 ( with the error being the internal uncertainty of the mass estimation code ) inside a radius @xmath53 ( @xmath54 kpc ) and @xmath151 inside @xmath131(@xmath152 kpc ) .
this estimate is a lower mass limit ( see [ ln : massestimate ] ) and assumes that the outer globular clusters are not influenced by the nearby giant s0 ngc 3115 . using two mass estimators ( i.e. pme and vme , see sect .
[ ln : massestimate ] ) and various assumptions for the systemic velocity and the cluster orbits , we derive masses between @xmath153m@xmath3 and @xmath5 m@xmath3 .
the mass increases with radius ( see table [ tab : massrad ] ) . inside r@xmath154
pc the mass - to - light ratio was found to be m / l@xmath155 @xcite and increases with radius , leading to m / l@xmath63 at @xmath23 10 kpc ( using the pme and assuming isotropic orbits ) , compatible with dark matter dominated outer regions .
however , we can not at present exclude the possibility that the high velocity dispersion is due to stripping of the two outer clusters by the nearby giant companion .
a kinematic analysis shows that the globular cluster system has a marginal net rotation of @xmath76 km s@xmath20 with a position angle of the rotation axis @xmath156 .
subtracting the net rotation we find a line - of - sight velocity dispersion of the globular cluster system of @xmath78 km s@xmath20 for the total sample of 7 globular clusters and @xmath79 km s@xmath20 for the inner 5 clusters ( see sect . [ ln : rotdisp ] ) .
we measure mean abundances ( using lick / ids passband definitions ) from a combined mean spectrum of all 7 globular clusters and derive a mean gcs metallicity of @xmath81[fe / h]@xmath82 dex .
all abundance ratios appear similar to the ones measured in milky way , m31 and ngc 1399 globular clusters . applying the color - metallicity calibration of @xcite
we obtain a photometric mean metallicity @xmath81[fe / h]@xmath157 dex with a dispersion of @xmath158})_{\rm gcs}=0.41\pm0.20 $ ] dex ( the error being the photometric uncertainty ) .
some 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 .
the observatory was made possible by the generous financial support of the w.m .
keck foundation .
allen , c. w. 1973 , astrophysical quantities , the athlone press ashman , k. m. , bird , c. m. , and zepf , s. e. 1994 , , 108 , 23482361 ashman , k. m. , and zepf , s. e. 1998 , globular cluster systems , cambridge university press bahcall , j. n. , and tremaine , s. 1981 , , 244 , 805819 bertin , e. , and arnouts , s. 1996 , a&as , 117 , 393404 binney , j. , and tremaine , s. 1994 , galactic dynamics , princeton series in astrophysics , princeton university press brodie , j. p. , and huchra , j. p. 1990
, , 362 , 503521 capaccioli , m. , cappellaro , e. , held , e. v. , and vietri , m. 1993 , a&a , 274 , 6978 ct , p. , marzke , r. o. , and west , m. j. 1998 , , 501 , 554 couture , j. , harris , w. e. , and allwright , j. w. b. 1990 , , 73 , 671683 da costa , g. s. , and armandroff , t. e. 1995 , , 109 , 2533 + dean , j. f. , warren , p. r. , and cousins , a. w. j. 1978 , , 183 , 569583 dekel , a. , and silk , j. 1986 , , 303 , 3955 durrell , p. r. , harris , w. e. , geisler , d. , and pudritz , r. e. 1996a , , 112 , 972 durrell , p. r. , mclaughlin , d. e. , harris , w. e. , and hanes , d. a. 1996b , , 463 , 543554 elson , r. a. w. , and freeman , k. c. 1985 , , 288 , 521530 ferguson , h. c. , and binggeli , b. 1994 , , 6 , 67122 gonzlez , j. j. 1993 , phd thesis , uco / lick observatory grillmair , c. j. , ajhar , e. a. , faber , s. m. , baum , w. a. , holtzman , j. a. , lauer , t. r. , lynds , c. r. , and oneil , e. j. jr .
1996 , , 111 , 2293 + harris , w. e. 1999 , catalog of parameters for milky way globular clusters , http://physun.physics.mcmaster.ca/globular.html heisler , j. , tremaine , s. , and bahcall , j. n. 1985 , , 298 , 817 hilker , m. , infante , l. , and richtler , t. 1999 , a&as , 138 , 5570 holtzman , j. , burrows , c. j. , casertano , s. , hester , j. j. , trauger , j. t. , watson , a. m. , and worthey , g. 1995 , , 107 , 10651093 horne , k. 1986 , , 98 , 609617 huchra , j. , stauffer , j. , van speybroeck , l. 1982 , , 259 , l57l61 van de kamp , p. 1967 , principles of astrometry , a series of books in mathematics , eds .
r. a. rosenbaum , and g. p. johnson , w. h. freeman and company kavelaars , j. j. 1997 , phd thesis , queen s university , ontario king , i. 1962 , , 67 , 471485 kissler - patig , m. , forbes , d.a . , and minniti , d. 1998a , 298 , 1123 kissler - patig , m. , brodie , j. p. , schroder , l. l. , forbes , d. a. , grillmair , c. j. , and huchra , j. p. 1998b
, , 115 , 105120 kissler - patig , m. , grillmair , c. j. , meylan , g. , brodie , j. p. , minniti , d. , and goudfrooij , p. 1999
, , 117 , 12061218 krist , j. , and hook , r. 1997 , the tinytim v4.4 handbook , http://scivax.stsci.edu/~krist/tinytim.html kundu , a. , and whitmore , b. c. 1998 , , 116 , 28412853 kundu , a. , whitmore , b. c. , sparks , w. b. , macchetto , f. d. , zepf , s. e. , and ashman , k. m. 1999 , , 513 , 733751 mateo , m. l. 1998 , , 36 , 435506 miller , b. w. , lotz , j. m. , ferguson , h. c. , stiavelli , m. , and whitmore , b. c. 1998 , , 508 , l133l137 minniti , d. , meylan , g. , and kissler - patig , m. 1996 , a&a , 312 , 49 oke , j. b. , cohen , j. g. , carr , m. , cromer , j. , dingizian , a. , harris , f. h. , labrecque , s. , lucinio , r. , schaal , w. , epps , h. , and miller , j. 1995 , pasp , 107 , 375385 peterson , r. c. , and caldwell , n. 1993 , , 105 , 14111419 pryor , c. & meylan , g. 1993 , in structure and dynamics of globular clusters , asp conf .
series vol .
djorgovski and g. meylan puzia , t. h. , kissler
patig , m. k. , brodie , j. p. , and huchra , j. p. 1999
, , 118 , 27342750 schlegel , d. j. , finkbeiner , d. p. , and davis , m. 1998 , , 500 , 525553 searle , l. , and zinn , r. 1978 , , 225 , 357 silverman , b. w. 1986 , density estimation for statistics and data analysis , chap and hall / crc press , inc .
trager , s. c. , king , i. r. , and djorgovski , s. 1995 , , 109 , 218241 trager , s. c. , worthey , g. , faber , s. m. , burstein , d. , and gonzlez , j. j. 1998 , , 116 , 1 + de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. g. jr . , buta , r. , paturel , g. , and fouque , p. 1991
, third reference catalogue of bright galaxies , berlin : springer worthey , g. , faber , s. m. , gonzlez , j. j. 1992 , , 398 , 6973 rr rl rl rl rlc & + + + + 1&l15 & 70492@xmath160 & 336 & 70386@xmath160 & 168 & & & 70407@xmath160 & 106 & @xmath161 + 2 & l3 & 63@xmath160 & 100 & 81@xmath160 & 136 & & & 69@xmath160 & 57 & @xmath162 + 3&l27 & 25223@xmath160 & 242 & 25242@xmath160 & 189 & & & 25235@xmath160 & 105 & @xmath161 + 4&l11 & 66@xmath160 & 113 & 105@xmath160 & 120 & & & 84@xmath160 & 58 & @xmath162 + 5&l18 & 194@xmath160 & 74 & 208@xmath160 & 97 & & & 199@xmath160 & 42 & @xmath162 + 6 & l7 & @xmath16350@xmath160 & 195 & @xmath16365@xmath160 & 141 & & & @xmath16360@xmath160 & 81 & @xmath162 + 7&l16&@xmath163211@xmath160 & 178 & @xmath163183@xmath160 & 162 & & & @xmath163196@xmath160 & 85 & @xmath162/ ? + 8 & d8 & 67@xmath160 & 127 & 101@xmath160 & 111 & & & 86@xmath160 & 59 & @xmath162 + 9&d21 & & & & & & & & & ? ? + 10&d10 & 100@xmath160 & 85 & 118@xmath160 & 73 & & & 110@xmath160 & 39 & @xmath162 + 11&d58 & 1864@xmath160 & 153 & 23692@xmath160 & 225 & & & 8766@xmath160 & 89 & @xmath161/ ? ?
+ 12&d73 & 4987@xmath160 & 307 & 5030@xmath160 & 225 & & & 5015@xmath160 & 128 & @xmath161/ ? ?
+ 13&d20 & 22542@xmath160 & 226 & 22576@xmath160 & 192 & & & 22562@xmath160 & 103 & @xmath161 + 14&d24 & 160@xmath160 & 224 & 154@xmath160 & 223 & & & 157@xmath160 & 112 & @xmath162 + 15&d14 & 455@xmath160 & 156 & 477@xmath160 & 149 & 478@xmath160 & 190 & 468@xmath160 & 71 & @xmath164/ ?
+ 16 & d6 & 17@xmath160 & 61 & 32@xmath160 & 66 & & & 24@xmath160 & 32 & @xmath162 + 17 & d1 & 143@xmath160 & 73 & 152@xmath160 & 83 & & & 147@xmath160 & 39 & @xmath162 + 18 & d5 & 124@xmath160 & 70 & 139@xmath160 & 73 & & & 131@xmath160 & 36 & @xmath162 + 19&l37 & 24531@xmath160 & 180 & 24614@xmath160 & 243 & & & 24560@xmath160 & 102 & @xmath161/ ? ?
+ 20&l44 & 22979@xmath160 & 210 & 23020@xmath160 & 223 & & & 22998@xmath160 & 108 & @xmath161/ ? ?
+ 21&l14 & 34@xmath160 & 111 & 67@xmath160 & 103 & & & 52@xmath160 & 53 & @xmath162 + 22&l51 & 26@xmath160 & 187 & 71@xmath160 & 196 & & & 47@xmath160 & 96 & @xmath162 + 23 & l1 & 386@xmath160 & 76 & 402@xmath160 & 79 & 453@xmath160 & 44 & 420@xmath160 & 29 & @xmath164 + 24&l63 & 545@xmath160 & 227 & 506@xmath160 & 222 & 668@xmath160 & 99 & 605@xmath160 & 74 & @xmath164/ ? + + + + 1 & l8 & 99434@xmath160 & 137 & 99453@xmath160 & 125 & & & 99444@xmath160 & 65 & @xmath161 + 2 & l4 & 55@xmath160 & 136 & 45@xmath160 & 109 & & & 49@xmath160 & 60 & @xmath162 + 3 & l6 & & & & & & & & & ? ?
+ 4&l23&@xmath163337@xmath160 & 227 & @xmath163259@xmath160 & 201 & & & @xmath163293@xmath160 & 106 & @xmath162/ ?
+ 5 & d9 & 9@xmath160 & 128 & 36@xmath160 & 117 & & & 24@xmath160 & 61 & @xmath162 + 6&d36 & @xmath16323@xmath160 & 152 & 13@xmath160 & 140 & & & @xmath1634@xmath160 & 73 & @xmath162 + 7&d26 & @xmath165 & 241 & @xmath1632545@xmath160 & 201 & & & @xmath1632567@xmath160 & 109 & @xmath162/ ? ?
+ 8 & d2 & @xmath16337@xmath160 & 111 & 24@xmath160 & 129 & & & @xmath16311@xmath160 & 59 & @xmath162 + 9&d25 & 540@xmath160 & 146 & 512@xmath160 & 158 & 540@xmath160 & 100 & 532@xmath160 & 60 & @xmath164 + 10&d46 & 628@xmath160 & 215 & 619@xmath160 & 208 & & & 623@xmath160 & 106 & @xmath164/ ? ?
+ 11 & d3 & @xmath16322@xmath160 & 113 & 12@xmath160 & 112 & & & @xmath1635@xmath160 & 56 & @xmath162 + 12&d15 & 677@xmath160 & 67 & 692@xmath160 & 72 & 672@xmath160 & 62 & 681@xmath160 & 30 & @xmath164 + 13 & d7 & 685@xmath160 & 70 & 706@xmath160 & 74 & 710@xmath160 & 34 & 703@xmath160 & 25 & @xmath164 + 14&d42 & 42007@xmath160 & 264 & 70422@xmath160 & 231 & & & 58100@xmath160 & 123 & @xmath161/ ? + 15&l36 & 59943@xmath160 & 215 & 3899@xmath160 & 176 & & & 26386@xmath160 & 96 & @xmath161/ ? ?
+ 16&d41 & 262@xmath160 & 206 & 376@xmath160 & 224 & 307@xmath160 & 50 & 308@xmath160 & 45 & @xmath162/ ? + 17&l37 & 43855@xmath160 & 162 & 17383@xmath160 & 168 & & & 31100@xmath160 & 82 & @xmath161/ ? ?
+ 18&l26 & 22836@xmath160 & 339 & 22952@xmath160 & 333 & & & 22895@xmath160 & 168 & @xmath161 + 19&l39 & 27415@xmath160 & 366 & 22664@xmath160 & 251 & & & 24184@xmath160 & 146 & @xmath161/ ? ?
+ 20&l52 & 87@xmath160 & 287 & 168@xmath160 & 217 & & & 139@xmath160 & 122 & @xmath162/ ? ?
+ 21&l17&131653@xmath160 & 187 & 131659@xmath160 & 79 & & & 131658@xmath160 & 51 & @xmath161 + 22&l38 & 24447@xmath160 & 135 & 26512@xmath160 & 335 & & & 24735@xmath160 & 89 & @xmath161/ ?
+ lcccc vme & 2.11 & & 0.04 & @xmath166 + pme@xmath167 & 4.77 & 2.28 & 0.16 & @xmath168 + pme@xmath169 & 9.54 & 5.05 & 0.32 & @xmath170 + pme@xmath171 & 7.15 & 2.77 & 0.24 & @xmath172 + pme@xmath173 & 3.18 & 1.41 & 0.11 & @xmath174 + llcccc @xmath175 & @xmath176 kpc & 4.77 & 2.28 & 0.16 & 7 + @xmath177 & @xmath178 kpc & 4.55 & 2.35 & 0.25 & 6 + @xmath179 & @xmath180 kpc & 1.82 & 1.03 & 0.08 & 5 + @xmath181 & @xmath182 kpc & 1.61 & 1.02 & 0.06 & 4 + lrrcrrc cnb & @xmath183 & @xmath184 & 0.195 & @xmath185 & @xmath186 & 0.197 + h+k & @xmath187 & @xmath188 & 0.155 & @xmath189 & @xmath190 & 0.242 + cnr & @xmath191 & @xmath192 & 0.328 & @xmath193 & @xmath194 & 0.474 + ch = gband&@xmath195 & @xmath196 & 0.126 & @xmath197 & @xmath198 & 0.246 + h@xmath85&@xmath199 & @xmath200&&@xmath201 & @xmath202 & + mgh & @xmath203 & @xmath204 & 0.257 & @xmath205 & @xmath206 & 0.372 + mgg & @xmath207 & @xmath208&&@xmath209 & @xmath210 & + mg2 & @xmath211 & @xmath212 & 0.496 & @xmath213 & @xmath214 & 0.698 + mgb & @xmath215 & @xmath216&&@xmath217 & @xmath218 & + fe52 & @xmath219 & @xmath220 & 0.146 & @xmath221 & @xmath222 & 0.221 + @xmath81[fe / h]@xmath223 & & + + & & + + cnb & @xmath224 & @xmath225 & 0.619 & @xmath226 & @xmath227 & 0.745 + h+k & @xmath228 & @xmath229 & 0.620 & @xmath230 & @xmath231 & 0.777 + cnr & @xmath232 & @xmath233 & 0.898 & @xmath234 & @xmath235 & 1.242 + ch = gband & @xmath236 & @xmath237 & 0.494 & @xmath238 & @xmath239 & 0.756 + h@xmath85 & @xmath240 & @xmath241&&@xmath242 & @xmath243 & + mgh & @xmath244 & @xmath245 & 0.726 & @xmath246 & @xmath247 & 1.050 + mgg & @xmath248 & @xmath249&&@xmath250 & @xmath202 & + mg2 & @xmath251 & @xmath252 & 1.295 & @xmath253 & @xmath254 & 1.796 + mgb & @xmath255 & @xmath256&&@xmath257 & @xmath258 & + fe52 & @xmath259 & @xmath260 & 0.467 & @xmath261 & @xmath262 & 0.701 + @xmath81[fe / h]@xmath223 & & + + & & + + & & & & & & + cnb & @xmath263 & @xmath264 & 0.419 & @xmath265 & @xmath266 & 0.854 + h+k & @xmath267 & @xmath268 & 0.399 & @xmath269 & @xmath270 & 0.791 + cnr & @xmath271 & @xmath272 & 0.660 & @xmath273 & @xmath274 & 1.323 + ch = gband & @xmath275 & @xmath276 & 0.363 & @xmath277 & @xmath278 & 0.818 + h@xmath85 & @xmath279 & @xmath280&&@xmath281 & @xmath282 & + mgh & @xmath283 & @xmath284 & 0.506 & @xmath285 & @xmath286 & 1.249 + mgg & @xmath287 & @xmath288&&@xmath289 & @xmath290 & + mg2 & @xmath291 & @xmath292 & 0.944 & @xmath293 & @xmath294 & 2.074 + mgb & @xmath295 & @xmath296&&@xmath297 & @xmath298 & + fe52 & @xmath299 & @xmath300 & 0.308 & @xmath301 & @xmath302 & 0.847 + @xmath81[fe / h]@xmath223 & & + + & & + + & & & & & & + cnb & @xmath303 & @xmath304 & 0.040 & @xmath305&@xmath306 & 1.215 + h+k & @xmath307 & @xmath308 & 0.159 & @xmath309&@xmath310 & 1.147 + cnr & @xmath311 & @xmath312 & 0.273 & @xmath313&@xmath314 & 1.608 + ch = gband & @xmath315 & @xmath316 & 0.139 & @xmath317&@xmath318 & 1.104 + h@xmath85 & @xmath319 & @xmath320 & & @xmath321&@xmath322 & + mgh & @xmath323 & @xmath324 & 0.223 & @xmath325&@xmath326 & 1.401 + mgg & @xmath327 & @xmath328 & & @xmath329&@xmath330 & + mg2 & @xmath331 & @xmath332 & 0.402 & @xmath333&@xmath334 & 2.275 + mgb & @xmath335 & @xmath336 & & @xmath337&@xmath338 & + fe52 & @xmath339 & @xmath340 & 0.146 & @xmath341&@xmath342 & 0.955 + @xmath81[fe / h]@xmath223 & & + lcc d7 & @xmath3430.23 & @xmath3440.20 + d14 & @xmath3450.39 & @xmath3460.20 + d15 & @xmath3470.20 & @xmath3480.20 + d25 & @xmath3480.37 & @xmath3490.30 + d46 & @xmath3500.82 & @xmath3510.45 + l1 & @xmath3520.19 & + l63 & @xmath3530.81 & + + & @xmath81[fe / h]@xmath354 & @xmath81[fe / h]@xmath355 + + rcccccrrrp0.005cmr + + + l15 & 151.46547275 & @xmath1638.02573829 & & & & 215.2 & 11474 & 70407&@xmath160&106 + l3 & 151.44513389 & @xmath1638.03275119 & & & & 197.4 & 10525 & 69&@xmath160 & 57 + l27 & 151.45595107 & @xmath1638.01588459 & & & & 166.1 & 8856 & 25235&@xmath160&105 + l11 & 151.47453033 & @xmath1638.00005692 & & & & 190.1 & 10139 & 84&@xmath160 & 58 + l18 & 151.45373013 & @xmath1638.00788268 & 21.65 @xmath160 0.05 & & 0.74 @xmath160 0.06 & 140.2 & 7475 & 199&@xmath160 & 42 + l7 & 151.43413151 & @xmath1638.01463921 & & & & 122.8 & 6549 & @xmath16360&@xmath160 & 81 + l16 & 151.43857103 & @xmath1638.00897480 & & & & 109.7 & 5850 & @xmath163196&@xmath160 & 85 + d8 & 151.44176680 & @xmath1638.00230439 & 21.70 @xmath160 0.02 & & 1.47 @xmath160 0.06 & 96.0 & 5117 & 86&@xmath160 & 59 + d21 & 151.42597410 & @xmath1638.00187100 & 21.64 @xmath160 0.03 & 1.37 @xmath160 0.06 & 0.70 @xmath160 0.04 & 72.1 & 3845 & + d10 & 151.45237834 & @xmath1637.98267887 & 20.97 @xmath160 0.02 & & 1.10 @xmath160 0.03 & 99.6 & 5312 & 110&@xmath160 & 39 + d58 & 151.43598000 & @xmath1637.99663820 & 23.22 @xmath160 0.11 & 2.09 @xmath160 0.10 & 1.09 @xmath160 0.18 & 67.1 & 3579 & 8766&@xmath160 & 89 + d73 & 151.43584940 & @xmath1637.98355023 & 23.38 @xmath160 0.09 & & 0.64 @xmath160 0.15 & 41.1 & 2190 & 5015&@xmath160&128 + d20 & 151.42182290 & @xmath1637.98467580 & 21.72 @xmath160 0.04 & 1.55 @xmath160 0.15 & 0.65 @xmath160 0.06 & 13.7 & 731 & 22562&@xmath160&103 + d24 & 151.43774990 & @xmath1637.97262110 & 21.78 @xmath160 0.04 & 1.28 @xmath160 0.07 & 0.67 @xmath160 0.07 & 58.0 & 3093 & 157&@xmath160&112 + d14 & 151.42099595 & @xmath1637.97540485 & 21.34 @xmath160 0.03 & & 0.90 @xmath160 0.04 & 26.4 & 1409 & 468&@xmath160 & 71 + d6 & 151.43119970 & @xmath1637.96549650 & 20.28 @xmath160 0.02 & 1.58 @xmath160 0.03 & 0.50 @xmath160 0.03 & 63.8 & 3401 & 24&@xmath160 & 32 + d1 & 151.40556550 & @xmath1637.97613822 & 18.85 @xmath160 0.01 & & 0.73 @xmath160 0.01 & 70.5 & 3757 & 147&@xmath160 & 39 + d5 & 151.41731295 & @xmath1637.96373779 & 20.32 @xmath160 0.01 & & 0.60 @xmath160 0.02 & 70.2 & 3742 & 131&@xmath160 & 36 + l37 & 151.42092628 & @xmath1637.95714247 & & & & 90.0 & 4800 & 24560&@xmath160&102 + l44 & 151.43991405 & @xmath1637.94358566 & & & & 148.5 & 7922 & 22998&@xmath160&108 + l14 & 151.43728995 & @xmath1637.93698967 & & & & 168.0 & 8961 & 52&@xmath160 & 53 + l51 & 151.42385948 & @xmath1637.93980723 & & & & 151.5 & 8082 & 47&@xmath160 & 96 + l1 & 151.41836728 & @xmath1637.93746430 & & & & 161.4 & 8609 & 420&@xmath160 & 29 + l63 & 151.40348832 & @xmath1637.93354949 & & & & 189.4 & 10102 & 605&@xmath160 & 74 + + + + l8 & 151.42908177 & @xmath1638.03330443 & & & & 185.8 & 9908 & 99444&@xmath160 & 65 + l4 & 151.45874258 & @xmath1638.02204502 & & & & 189.3 & 10095 & 49&@xmath160 & 60 + l6 & 151.43119125 & @xmath1638.02568813 & & & & 159.5 & 8504 & + l23 & 151.44361588 & @xmath1638.01586187 & & & & 140.1 & 7469 & @xmath163293&@xmath160&106 + d9 & 151.43313299 & @xmath1638.00698070 & 21.63 @xmath160 0.02 & & 1.52 @xmath160 0.06 & 95.5 & 5090 & 24&@xmath160 & 61 + d36 & 151.44962050 & @xmath1637.99643065 & 22.86 @xmath160 0.07 & & 1.70 @xmath160 0.17 & 103.9 & 5539 & @xmath1634&@xmath160 & 73 + d26 & 151.44384880 & @xmath1637.98342710 & 22.49 @xmath160 0.07 & 1.25 @xmath160 0.11 & 1.43
@xmath160 0.13 & 69.4 & 3700 & @xmath1632567&@xmath160&109 + d2 & 151.43874410 & @xmath1637.99131620 & 20.32 @xmath160 0.02 & 2.69 @xmath160 0.04 & 1.60 @xmath160 0.03 & 61.2 & 3264 & @xmath16311&@xmath160 & 59 + d25 & 151.43899620 & @xmath1637.98802330 & 21.78 @xmath160 0.04 & 1.27 @xmath160 0.07 & 0.69 @xmath160 0.06 & 56.3 & 3005 & 532&@xmath160 & 60 + d46 & 151.42821583 & @xmath1637.98176333 & 22.53 @xmath160 0.05 & & 0.74 @xmath160 0.09 & 13.4 & 716 & 623&@xmath160&106 + d3 & 151.44378760 & @xmath1637.97441980 & 20.74 @xmath160 0.02 & 3.32 @xmath160 0.03 & 1.53 @xmath160 0.04 & 74.0 & 3947 & @xmath1635&@xmath160 & 56 + d15 & 151.43225130 & @xmath1637.97337860 & 21.30 @xmath160 0.02 & 1.52 @xmath160 0.05 & 0.80 @xmath160 0.04 & 41.4 & 2208 & 681&@xmath160 & 30 + d7 & 151.41813758 & @xmath1637.97210601 & 20.53 @xmath160 0.03 & & 0.89 @xmath160 0.04 & 41.8 & 2231 & 703&@xmath160 & 25 + d42 & 151.44129540 & @xmath1637.96256740 & 22.79 @xmath160 0.07 & 1.56 @xmath160 0.10 & 1.63 @xmath160 0.18 & 91.9 & 4902 & 58100&@xmath160&123 + l36 & 151.45635892 & @xmath1637.95500702 & & & & 149.4 & 7966 & 26386&@xmath160 & 96 + d41 & 151.40979676 & @xmath1637.96307913 & 22.19 @xmath160 0.04 & & 0.49 @xmath160 0.06 & 85.6 & 4563 & 308&@xmath160 & 45 + l37 & 151.42092628 & @xmath1637.95714247 & & & & 90.0 & 4800 & 31100&@xmath160 & 82 + l26 & 151.44282349 & @xmath1637.94889087 & & & & 135.7 & 7236 & 22895&@xmath160&168 + l39 & 151.42936599 & @xmath1637.94366459 & & & & 138.8 & 7400 & 24184&@xmath160&146 + l52 & 151.41051686 & @xmath1637.94476283 & & & & 142.6 & 7606 & 139&@xmath160&122 + l17 & 151.45223278 & @xmath1637.92867126 & & & & 215.7 & 11504 & 131658&@xmath160 & 51 + l38 & 151.42761022 & @xmath1637.93031858 & & & & 186.0 & 9921 & 24735&@xmath160 & 89 + | the properties of globular clusters in dwarf galaxies are key to understanding the formation of globular cluster systems , and in particular in verifying scenarios in which globular cluster systems of larger galaxies formed ( at least partly ) from the accretion of dwarf galaxies . here
, we revisit the globular cluster system of the de , n galaxy ngc 3115 dw1 a companion of the nearby s0 galaxy ngc 3115 adding keck / lris spectroscopy and hst / wfpc2 imaging to previous ground - based photometry .
spectra for seven globular clusters reveal normal abundance ratios with respect to the milky way and m31 clusters , as well as a relatively high mean metallicity ( [ fe / h]@xmath0 dex ) .
crude kinematics indicate a high velocity dispersion within 10 kpc which could either be caused by dark matter dominated outer regions , or by the stripping of outer globular clusters by the nearby giant galaxy ngc 3115 .
the total galaxy mass out to 3 and 10 kpc lies between @xmath1 and @xmath2m@xmath3 and @xmath4 and @xmath5 m@xmath3 , respectively , depending on the mass estimator used and the assumptions on cluster orbits and systemic velocity .
the hst imaging allows measurement of sizes for two clusters , returning core radii around 2.0 pc , similar to the sizes observed in other galaxies .
spectroscopy allows an estimate of the degree of contamination by foreground stars or background galaxies for the previous ground - based photometry , but does not require a revision of previous results : ngc 3115 dw1 hosts around @xmath6 clusters which corresponds to a specific frequency of @xmath7 , on the high end for massive des .
given its absolute magnitude ( @xmath8 mag ) and the properties of its cluster system , ngc 3115 dw1 appears to be a transition between a luminous de and low - luminosity e galaxy . |
in 1989 witten derived the jones polynomial from quantum field theory based on the chern - simon lagrangian @xcite@xcite .
inspired by witten s work in this paper we shall derive a way to construct quantum knots and knot invariant from a quantum gauge model of electrodynamics and its nonabelian generalization . from this quantum gauge model
we derive a conformal field theory which includes the kac - moody algebra and the knizhnik - zamolodchikov equation @xcite . here as a difference from the usual conformal field theory we can derive two quantum knizhnik - zamolodchikov ( kz ) equations which are dual to each other .
these two quantum kz equations are equations for the product of @xmath0 wilson lines @xmath1 which are defined by the gauge model . these two quantum kz equations can be regarded as a quantum yang - mills equation since it is analogous to the classical yang - mills equation derived from the classical yang - mills gauge model . from the two quantum kz equations we derive the skein relation of the homfly polynomial @xcite-@xcite .
in this derivation we represent the uppercrossing , zero crossing and undercrossing of two pieces of curves by the products of two quantum wilson lines @xmath2 and @xmath3 .
then by the braiding of quantum wilson lines we construct generalized wilson loops .
we show that these generalized wilson loops can represent knot and link diagrams and thus can be defined as quantum knots and links . from quantum knots and links
we derive quantum invariant of knots and links .
we show that this invariant gives a classification of knots and links . from this invariant
we then derive quantum invariant of closed 3-manifolds .
we show that this quantum invariant of closed 3-manifolds give a classification of closed manifolds . from this classification
we then prove the poincar conjecture .
this paper is organized as follows . in section 2
we give a brief description of a quantum gauge model of electrodynamics and its nonabelian generalization . in this paper
we shall consider a nonabelian generalization with a @xmath4 gauge symmetry . in section 3
we define the classical wilson loop . in section 4
we derive the defintion of the generator of the wilson line . from this definition
we derive a conformal field theory which includes the affine kac - moody algebra , the virasoro energy operator ( or the virasoro energy - momentum tensor ) and the virasoro algebra . in section 5
we derive the quantum kz equation in dual form . in section 6
we compute the solutions of the quantum kz equation in dual form . in section 7
we compute the quantum wilson line . in section 8
we represent the braiding of two pieces of curves by defining the braiding of two quantum wilson lines . in section 9
we derive the skein relation for the homfly polynomial . in section 10
we compute the quantum wilson loop . in section 11
we define generalized wilson loops which will be as quantum knots . in section 12
we give some examples of generalized wilson loops and show that they have the properties of the corresponding knot diagram and thus may be regarded as quantum knots . in section 13
we show that this generalized wilson loop is a complete copy of the corresponding knot diagram and thus we may call a generalized wilson loop as a quantum knot . from quantum knots
we have a knot invariant of the form @xmath5 where @xmath6 denotes a wilson loop and @xmath7 is the braiding matrix and is the monodromy of the quantum kz equation and @xmath8 is an integer . we show that this knot invariant classifies knots and that knots can be one - to - one assigned with the integer @xmath8 . in section 14
we give more computations of quantum knots and their knot invariant .
then in section 15 and 16 with the integer @xmath8 we give a classification table of knots where we show that prime knots ( and only prime knots ) are assigned with prime integer @xmath8 . in section 17
we give examples of invariant of links . in section 18
we give a classification of links . in section 19
we construct a quantum invariant of 3-manifolds .
we first construct quantum invariant of closed three - manifolds obtained by dehn surgery on framed knots .
we then introduce the concept of minimal link to construct quantum invariant of closed three - manifolds obtained by dehn surgery on framed links . then by using the lickorish - wallace theorem which states that any closed 3-manifold @xmath9 can be obtained from a dehn surgery on a framed link @xmath10
we show that this invariant gives a classification of closed 3-manifolds .
then in section 20 we use this classification to prove the poincar conjecture .
we shall first establish a quantum gauge model .
this quantum gauge model will be as a physical motivation for introducing operators which will be called wilson loop and wilsom line as analogous to the wilson loops in the existing quantum field theories .
then the definition of wilson loop and wilson line and the definition of a generator @xmath11 of the wilson line will be as the basis of the mathematical foundation of this paper ( in order to simplify the mathematics of this paper we treat this quantum gauge model as a physical motivation instead of as the mathematical foundation of this paper ) . we shall show that the generator @xmath11 gives an affine kac - moody algebra and a virasoro energy operator @xmath12 with central charge @xmath13 . from @xmath11 and @xmath12 we shall derive the quantum kz equation in dual form which will be regarded as the quantum yang - mills equation . from this quantum kz equation
we then construct generalized wilson loops which will be as quantum knots and links .
let us construct a quantum gauge model , as follows . in probability theory
we have the wiener measure @xmath14 which is a measure on the space @xmath15 $ ] of continuous functions @xcite .
this measure is a well defined mathematical theory for the brownian motion and it may be symbolically written in the following form : @xmath16 where @xmath17 is the energy integral of the brownian particle and @xmath18 is symbolically a product of lebesgue measures @xmath19 and @xmath20 is a normalized constant .
once the wiener measure is defined we may then define other measures on @xmath15 $ ] as follows@xcite .
let a potential term @xmath21 be added to @xmath22 .
then we have a measure @xmath23 on @xmath15 $ ] defined by : @xmath24 under some condition on @xmath25 we have that @xmath23 is well defined on @xmath15 $ ] .
let us call ( [ wiener2 ] ) as the feymann - kac formula @xcite .
let us then follow this formula to construct a quantum model of electrodynamics , as follows . then similar to the formula ( [ wiener2 ] ) we construct a quantum model of electrodynamics from the following energy integral : @xmath26ds \label{1.1}\ ] ] where the complex variable @xmath27 and the real variables @xmath28 and @xmath29 are continuous functions in a form that they are in terms of an arbitrary ( continuously differentiable ) closed curve @xmath30 in the complex plane where @xmath31 is a parameter representing the proper time in relativity ( we shall also write @xmath32 in the complex variable form @xmath33 ) . the complex variable @xmath27 represents a field of matter ( such as the electron ) ( @xmath34 denotes its complex conjugate ) and the real variables @xmath28 and @xmath29 represent a connection ( or the gauge field of the photon ) and @xmath35 denotes the electric charge .
the integral ( [ 1.1 ] ) has the following gauge symmetry : @xmath36 where @xmath37 is a continuously differentiable real - valued function of @xmath38 .
we remark that this model is similar to the usual yang - mills gauge model .
a feature of ( [ 1.1 ] ) is that it is not formulated with the four - dimensional space - time but is formulated with the one dimensional proper time .
this one dimensional nature let this model avoid the usual utraviolet divergence difficulty of quantum fields .
similar to the usual yang - mills gauge theory we can generalize this gauge model with @xmath39 gauge symmetry to nonabelian gauge models . as an illustration
let us consider @xmath4 gauge symmetry .
similar to ( [ 1.1 ] ) we consider the following energy integral : @xmath40ds \label{n1}\ ] ] where @xmath41 is a two dimensional complex vector ; @xmath42 @xmath43 where @xmath44 denotes a component of a gauge field @xmath45 ; @xmath46 denotes a generator of @xmath4 ( here for simplicity we choose a convention that the complex @xmath47 is absorbed by @xmath46 ) ; and @xmath48 , @xmath43 where @xmath49 denotes the charge of interaction ( for simplicity let us set @xmath50 ) . from ( [ n1 ] ) we can develop a nonabelian gauge model as similar to that for the above abelian gauge model .
we have that ( [ n1 ] ) is invariant under the following gauge transformation : @xmath51 where @xmath52 and @xmath53 .
we shall mainly consider the case that @xmath54 is a function of the form @xmath55 where @xmath56 are analytic functions of @xmath38 ( we let @xmath57 and we write @xmath58 ) .
the above gauge model is based on the banach space @xmath59 of continuous functions @xmath60 on the one dimensional interval @xmath61 .
since @xmath10 is positive and the model is one dimensional ( and thus is simpler than the usual two dimensional yang - mills gauge model ) we have that this gauge model is similar to the wiener measure except that this gauge model has a gauge symmetry .
this gauge symmetry gives a degenerate degree of freedom . in the physics literature the usual way to treat the degenerate degree of freedom of gauge symmetry is to introduce a gauge fixing condition to eliminate the degenerate degree of freedom where each gauge fixing will give equivalent physical results@xcite .
there are various gauge fixing conditions such as the lorentz gauge condition , the feynman gauge condition , etc .
we shall later in section [ sec6 ] ( on the kac - moody algebra ) adopt a gauge fixing condition for the above gauge model .
this gauge fixing condition will also be used to derive the quantum kz equation in dual form which will be regarded as a quantum yang - mill equation since its role will be similar to the classical yang - mill equation derived from the classical yang - mill gauge model .
since @xmath10 is positive we have that without gauge fixing condition the above gauge model is a positive linear functional on the banach space @xmath62 of continuous functions on @xmath59 and is multivalued in the sense that each gauge fixing gives a value .
* remark*. in this paper the main aim of introducing this quantum gauge model is to derive the quantum kz equation in dual form which will be regarded as a quantum yang - mills equation ( or as a quantum euler - lagrange equation ) . from this quantum kz equation in dual form
we then construct quantum knots and links . from quantum knots and links
we then prove the poincare conjecture .
similar to the wilson loop in quantum field theory @xcite from our quantum model we introduce an analogue of wilson loop , as follows .
* definition*. a classical wilson loop @xmath63 is defined by : @xmath64 where @xmath7 denotes a representation of @xmath4 ; @xmath65 is a fixed curve where the quantum gauge models are based on it as specified in the above section . as usual the notation @xmath66 in the definition of @xmath63 denotes a path - ordered product @xcite@xcite@xcite .
let us give some remarks on the above definition of wilson loop , as follows .
\1 ) we use the notation @xmath67 to mean the wilson loop @xmath63 which is based on the whole closed curve @xmath68 . here for convenience we use only the end points @xmath69 and @xmath70 of the curve @xmath68 to denote this wilson loop ( we keep in mind that the definition of @xmath67 depends on the whole curve @xmath68 connecting @xmath69 and @xmath70 ) .
then we extend the definition of @xmath63 to the case that @xmath68 is not a closed curve with @xmath71 .
when @xmath68 is not a closed curve we shall call @xmath67 as a wilson line .
\2 ) in constructing the wilson loop we need to choose a representation @xmath7 of the @xmath4 group .
we shall see that because a wilson line @xmath67 is with two variables @xmath69 and @xmath70 a natural representation of a wilson line or a wilson loop is the tensor product of the usual two dimensional representation of @xmath4 for constructing the wilson loop .
we first have the following theorem on @xmath73 : for a given continuous path @xmath74 on @xmath75 $ ] the wilson line @xmath73 exists on this path and has the following transition property : @xmath76 where @xmath73 denotes the wilson line of a curve @xmath68 which is with @xmath69 as the starting point and @xmath70 as the ending point and @xmath38 is a point on @xmath68 between @xmath69 and @xmath70 .
* proof*. we have that @xmath73 is a limit ( whenever exists ) of ordered product of @xmath77 and thus can be written in the following form : @xmath78 a_i(z(s_2))\frac{dx^i(s_2)}{ds}ds_2 + \cdot\cdot\cdot \end{array } \label{df3}\ ] ] where @xmath79 and @xmath80 . then since @xmath81 are continuous on @xmath82 $ ] and @xmath83 are continuously differentiable on @xmath84 $ ] we have that the series in ( [ df3 ] ) is absolutely convergent .
thus the wilson line @xmath73 exists .
then since @xmath73 is the limit of ordered product we can write @xmath73 in the form @xmath85 by dividing @xmath68 into two parts at @xmath38 .
this proves the theorem . @xmath72 * remark ( classical and quantum wilson loop)*. this theorm means that the wilson line @xmath73 exists in the classical pathwise sense where @xmath81 are as classical paths on @xmath75 $ ] .
this pathwise version of the wilson line @xmath73 ; from the feymann path integral point of view ; is as a partial description of the quantum version of the wilson line @xmath73 which is as an operator when @xmath81 are as operators .
we shall in the next section derive and define a quantum generator @xmath11 of @xmath73 from the quantum gauge model . then by using this generator @xmath11 we shall compute the quantum version of the wilson line @xmath73 .
we shall denote both the classical version and quantum version of wilson line by the same notation @xmath73 when there is no confusion .
@xmath72 * remark*. we remark again that in order to simplify the mathematics of this paper we treat the above quantum gauge model as a physical motivation instead of as the mathematical foundation of this paper .
the mathematical foundation of this paper will base on the definition of the wilson line @xmath73 and the generator @xmath11 instead of the above quantum gauge model which is as a physical motivation for introducing the wilson line @xmath73 and the generator @xmath11 .
@xmath72 by following the usual approach from a gauge transformation we have the following symmetry on wilson lines ( this symmetry is sometimes called the chiral symmetry ) @xcite : under an analytic gauge transformation with an analytic function @xmath86 we have the following symmetry : @xmath87 where @xmath88 is a wilson line with gauge field @xmath89 .
* proof*. let us prove this symmetry as follows .
let @xmath90 and @xmath91 where @xmath92 .
following kauffman @xcite we have @xmath93 from ( [ n5b ] ) we have that ( [ n5 ] ) holds since ( [ n5 ] ) is the limit of ordered product in which the left - side factor @xmath94 in ( [ n5b ] ) with @xmath95 is canceled by the right - side factor @xmath96 of ( [ n5b ] ) where @xmath97 with @xmath98 .
this proves the theorem .
@xmath72 as analogous to the wzw model in conformal field theory @xcite@xcite from the above symmetry we have the following formulas for the variations @xmath99 and @xmath100 with respect to this symmetry : @xmath101 and @xmath102 where @xmath38 and @xmath103 are independent variables and @xmath104 when @xmath105 . in ( [ k1 ] )
the variation is with respect to the @xmath38 variable while in ( [ k2 ] ) the variation is with respect to the @xmath103 variable .
this two - side - variations when @xmath106 can be derived as follows .
for the left variation we may let @xmath86 be analytic in a neighborhood of @xmath38 and continuous differentiably extended to a neighborhood of @xmath103 such that @xmath107 in this neighborhood of @xmath103 .
then from ( [ n5 ] ) we have that ( [ k1 ] ) holds .
similarly we may let @xmath108 be analytic in a neighborhood of @xmath103 and continuous differentiably extended to a neighborhood of @xmath38 such that @xmath109 in this neighborhood of @xmath38 .
then we have that ( [ k2 ] ) holds .
this section has two related purposes .
one purpose is to find a gauge fixing condition for eliminating the degenerate degree of freedom from the gauge invariance of the above quantum gauge model in section 2 .
then another purpose is to find an equation for defining a generator @xmath11 of the wilson line @xmath1 .
this defining equation of @xmath11 can then be used as a gauge fixing condition .
thus with this defining equation of @xmath11 the construction of the quantum gauge model in section 2 is then completed ( we remark that we shall let the definition of the wilson line and the definition of the generator @xmath11 as the mathematical foundation of this paper while the quantum gauge model is as a physical motivation for deriving these two definitions ) .
we shall derive a quantum loop algebra ( or the affine kac - moody algebra ) structure from the wilson line @xmath1 for the generator @xmath11 of @xmath1 . to this end
let us first consider the classical case . since @xmath1 is constructed from @xmath110 we have that the mapping @xmath111 ( we consider @xmath1 as a function of @xmath38 with @xmath103 being fixed ) has a loop group structure @xcite@xcite .
for a loop group we have the following generators : @xmath112 these generators satisfy the following algebra : @xmath113 = if_{abc}j_{m+n}^c \label{km2}\ ] ] this is the so called loop algebra @xcite@xcite .
let us then introduce the following generating function @xmath11 : @xmath114 where we define @xmath115 from @xmath11 we have @xmath116 where @xmath117 denotes a closed contour integral with center @xmath38 .
this formula can be interpreted as that @xmath11 is the generator of the loop group and that @xmath118 is the directional generator in the direction @xmath119 .
we may generalize @xmath120 to the following directional generator : @xmath121 where the analytic function @xmath122 is regarded as a direction and we define @xmath123 then since @xmath124 , from the variational formula ( [ km5 ] ) for the loop algebra of the loop group of @xmath4 we have that the variation of @xmath1 in the direction @xmath125 is given by @xmath126 now let us consider the quantum case which is based on the quantum gauge model in section 2 . for this quantum case we shall define a quantum generator @xmath11 which is analogous to the @xmath11 in ( [ km3 ] ) .
we shall choose the equations ( [ n8b ] ) and ( [ n6 ] ) as the equations for defining the quantum generator @xmath11 .
let us first give a formal derivation of the equation ( [ n8b ] ) , as follows .
let us consider the following formal functional integration : @xmath127 where @xmath128 denotes a field from the quantum gauge model ( we first let @xmath103 be fixed as a parameter ) .
let us do a calculus of variation on this integral to derive a variational equation by applying a gauge transformation on ( [ n8a ] ) as follows ( we remark that such variational equations are usually called the ward identity in the physics literature ) .
let @xmath129 be regarded as a coordinate system of the integral ( [ n8a ] ) . under a gauge transformation ( regarded as a change of coordinate ) with gauge function @xmath130 this coordinate
is changed to another coordinate denoted by @xmath131 .
as similar to the usual change of variable for integration we have that the integral ( [ n8a ] ) is unchanged under a change of variable and we have the following equality : @xmath132 where @xmath133 denotes the wilson line based on @xmath134 and @xmath135 and similarly @xmath136 denotes the field obtained from @xmath128 with @xmath137 replaced by @xmath138
. then it can be shown that the differential is unchanged under a gauge transformation @xcite : @xmath139 also by the gauge invariance property the factor @xmath140 is unchanged under a gauge transformation .
thus from ( [ int ] ) we have @xmath141 where the correlation notation @xmath142 denotes the integral with respect to the differential @xmath143 we can now carry out the calculus of variation . from the gauge transformation we have the formula @xmath144 ( @xmath58 ) .
this gauge transformation gives a variation of @xmath1 with the gauge function @xmath145 as the variational direction @xmath54 in the variational formulas ( [ km5 ] ) and ( [ km6 ] ) .
thus analogous to the variational formula ( [ km6 ] ) we have that the variation of @xmath1 under this gauge transformation is given by @xmath146 where the generator @xmath11 for this variation is to be specified .
this @xmath11 will be a quantum generator which generalizes the classical generator @xmath11 in ( [ km6 ] ) .
thus under a gauge transformation with gauge function @xmath145 from ( [ w1 ] ) we have the following variational equation : @xmath147\rangle \label{w2}\ ] ] where @xmath148 denotes the variation of the field @xmath128 in the direction @xmath145 . from this equation
an ansatz of @xmath11 is that @xmath11 satisfies the following equation : @xmath149 = 0 \label{n8bb}\ ] ] from this equation we have the following variational equation : @xmath150 this completes the formal calculus of variation . now
( with the above derivation as a guide ) we choose the following equation ( [ n8b ] ) as one of the equation for defining the generator @xmath11 : @xmath151 where we generalize the direction @xmath58 to the analytic direction @xmath152 ( this generalization has the effect of extending the real measure to include the complex feymann path integral ) .
let us now choose one more equation for determine the generator @xmath11 in ( [ n8b ] ) . this choice will be as a gauge fixing condition .
as analogous to the wzw model in conformal field theory @xcite@xcite @xcite let us consider a @xmath11 given by @xmath153 where we define @xmath154 and we set @xmath105 after the differentiation with respect to @xmath38 ; @xmath155 is a constant which is fixed when the @xmath11 is determined to be of the form ( [ n6 ] ) and the minus sign is chosen by convention . in the wzw model @xcite@xcite the @xmath11 of the form ( [ n6 ] ) is the generator of the chiral symmetry of the wzw model .
we can write the @xmath11 in ( [ n6 ] ) in the following form : @xmath156 we see that the generators @xmath157 of @xmath4 appear in this form of @xmath11 and this form is analogous to the classical @xmath11 in ( [ km3 ] ) .
this shows that this @xmath11 is a possible candidate for the generator @xmath11 in ( [ n8b ] ) .
since @xmath1 is constructed by gauge field we need to have a gauge fixing for the computations related to @xmath1
. then since the @xmath11 in ( [ n8b ] ) and ( [ n6 ] ) is constructed from @xmath1 we have that in defining this @xmath11 as the generator @xmath11 of @xmath1 we have chosen a condition for the gauge fixing . in this paper
we shall always choose this defining equations ( [ n8b ] ) and ( [ n6 ] ) for @xmath11 as the gauge fixing condition . in summary
we introduce the following definition .
* definition * the generator @xmath11 of the quantum wilson line @xmath1 whose classical version is defined by ( [ n4 ] ) , is an operator defined by the two conditions ( [ n8b ] ) and ( [ n6 ] ) . @xmath72 * remark*. we remark that the condition ( [ n6 ] ) first defines @xmath11 classically .
then the condition ( [ n8b ] ) raises this classical @xmath11 to the quantum generator @xmath11 .
@xmath72 now we want to show that this generator @xmath11 in ( [ n8b ] ) and ( [ n6 ] ) can be uniquely solved ( this means that the gauge fixing condition has already fixed the gauge that the degenerate degree of freedom of gauge invariance has been eliminated so that we can carry out computation ) . before solving @xmath11
we give the following remark .
* remark*. we remark again that in the above of this paper we have introduced a quantum gauge model as a physical motivation for introducing the wilson loop and wilson line defined by ( [ n4 ] ) and the generator @xmath11 defined by the two conditions ( [ n8b ] ) and ( [ n6 ] ) . in the following of this paper all the mathematics will be based on these two definitions .
thus we let these two definitions be as the mathematical foundation of this paper and treat the quantum gauge model as a physical motivation for deriving these two definitions .
let us now solve @xmath11 . from ( [ n5 ] ) and ( [ n6 ] ) we have that the variation @xmath158 of the generator @xmath11 in ( [ n6 ] ) is given by @xcite(p.622 )
@xcite : @xmath159 from ( [ n8b ] ) and ( [ n8c ] ) we have that @xmath11 satisfies the following relation of current algebra @xcite@xcite@xcite : @xmath160 where as a convention the regular term of the product @xmath161 is omitted .
then by following @xcite@xcite@xcite from ( [ n8d ] ) and ( [ km8 ] ) we can show that the @xmath118 in ( [ km3 ] ) for the corresponding laurent series of the quantum generator @xmath11 satisfy the following kac - moody algebra : @xmath113 = if_{abc}j_{m+n}^c + km\delta_{ab}\delta_{m+n , 0 } \label{n8}\ ] ] where @xmath162 is usually called the central extension or the level of the kac - moody algebra .
* remark*. let us also consider the other side of the chiral symmetry .
similar to the @xmath11 in ( [ n6 ] ) we define a generator @xmath163 by : @xmath164 where after differentiation with respect to @xmath103 we set @xmath165 .
let us then consider the following formal correlation : @xmath166 where @xmath38 is fixed . by an approach similar to the above derivation of ( [ n8b ] ) we have the following variational equation : @xmath167 where as a gauge fixing we choose the @xmath163 in ( [ n8b1 ] )
be the @xmath163 in ( [ d1 ] ) .
then similar to ( [ n8c ] ) we also have @xmath168 then from ( [ n8b1 ] ) and ( [ n8c1 ] ) we can derive the current algebra and the kac - moody algebra for @xmath163 which are of the same form of ( [ n8d ] ) and ( [ n8 ] ) .
from this we have @xmath169 . @xmath72
with the above current algebra @xmath11 and the formula ( [ n8b ] ) we can now follow the usual approach in conformal field theory to derive a quantum knizhnik - zamolodchikov ( kz ) equation for the product of primary fields in a conformal field theory @xcite@xcite@xcite .
we shall derive the kz equation for the product of @xmath0 wilson lines @xmath170 . here
an important point is that from the two sides of @xmath170 we can derive two quantum kz equations which are dual to each other .
these two quantum kz equations are different from the usual kz equation in that they are equations for the quantum operators @xmath170 while the usual kz equation is for the correlations of quantum operators . with this difference the following derivation of kz equation for deriving these two quantum kz equations
is well known in conformal field theory @xcite@xcite .
the reader may skip this derivation of kz equation and just look at the form of the virasoro energy operator @xmath171 ( which is usually called the virasoro energy - momentum tensor ) and the virasoro algebra and the form of these two quantum kz equations .
let us first consider ( [ k1 ] ) . from ( [ n8b ] ) and ( [ k1 ] )
we have @xmath172 where as a convention the regular term of the product @xmath173 is omitted .
following @xcite and @xcite let us define an energy operator @xmath171 by @xmath174 where @xmath49 is the dual coxter number of @xmath4 @xcite . in ( [ k4 ] )
the symbol @xmath175 denotes the normal ordering of the operator @xmath176 which can be defined as follows @xcite@xcite .
let a product of operators @xmath177 be written in the following laurent series form : @xmath178 the singular part of ( [ normal ] ) is called the contraction of @xmath177 and will be denoted by @xmath179 .
then the term @xmath180 is called the normal ordering of @xmath177 and we denote @xmath180 by @xmath181 .
these terms are originally from quantum field theory .
we remark that in @xcite the notation @xmath182 is used to generalize the original definition of @xmath183 for products of free fields . here for simplicity we shall always use the notation @xmath183 to mean the normal ordering of @xmath184 . from this definition of normal
ordering we have the following form of normal ordering @xcite : @xmath185 this form can be checked by taking the contour integral on the laurent series expansion of @xmath177 .
alternatively we may let ( [ conform1a ] ) be the definition of normal ordering .
we then define ( [ k4 ] ) by ( [ conform1a ] ) with @xmath186 .
the above definition of the energy operator @xmath171 is called the sugawara construction @xcite .
we first have the following well known theorem on @xmath171 in conformal field theory @xcite : the operator product @xmath187 is given by the following formula : @xmath188 for some constant @xmath189 ( @xmath190 for the group @xmath4 ) and as a convention we omit the regular term of this product .
* proof*. in @xcite there is a detail proof of this theorem . here we want to remark that the formula ( [ n8d ] ) for the product @xmath191 is used for the proof of this theorem .
@xmath72 from this theorem we then have the following virasoro algebra of the mode expansion of @xmath171 @xcite@xcite : let us write @xmath171 in the following laurent series form : @xmath192 this means that the modes @xmath193 are defined by @xmath194 then we have that @xmath195 form a virasoro algebra : @xmath196 = ( n - m)l_{n+m } + \frac{c}{12}n(n^2 - 1)\delta_{n+m,0 } \label{conform11}\ ] ] from the formula ( [ n8d ] ) for the product @xmath197 we have the following operator product expansion @xcite : @xmath198 then we have the following operator product of @xmath171 with an operator @xmath199 : @xmath200 from ( [ conform7 ] ) and ( [ conform13 ] ) we have that @xmath201 and @xmath202 .
thus we have @xmath203 on the other hand as shown in @xcite by using the laurent series expansion of @xmath204 in the section on kac - moody algebra we can compute the normal ordering @xmath175 from which we have the laurent series expansion of @xmath171 with @xmath205 given by @xcite : @xmath206 where since @xmath207 and @xmath208 commute each other the ordering of them is irrelevant . from ( [ conform15 ] )
we then have @xmath209 since @xmath210 for @xmath211 .
it follows from ( [ conform14 ] ) and ( [ conform16 ] ) that we have the following equality : @xmath212 then form ( [ k3 ] ) we have @xmath213 from ( [ k7 ] ) and ( [ k8 ] ) we then have @xmath214 now let us consider a product of @xmath0 wilson lines : @xmath215 . let this product be represented as a tensor product when @xmath216 and @xmath217 , @xmath218 are all independent variables .
then from ( [ k9 ] ) we have @xmath219 where the second equality is from the definition of tensor product for which we define @xmath220\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \label{k9aa}\ ] ] with this formula ( [ k9a ] ) we can now follow @xcite and @xcite to derive the kz equation . for a easy reference let us present this derivation in @xcite and @xcite as follows . from the laurent series of @xmath221 we have @xmath222 where the line integral is on a contour encircling @xmath216 .
we also let this contour encircles all other @xmath223 so that the effects from wilson lines @xmath224 for @xmath225 will all be counted
. then we have @xmath226\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \end{array } \label{norm3aa}\ ] ] where the second equality is from the @xmath227 product formula ( [ k3 ] ) .
then by a deformation of the contour integral in ( [ norm3aa ] ) into a sum of @xmath0 contour integrals such that each contour integral encircles one and only one @xmath223 we have : @xmath228\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \\ & \\ = & \sum_{j=1,j\neq i}^n \frac{1}{z_j - z_i}w(z_1 , z_1^{\prime})\cdot\cdot\cdot [ -t^a w(z_j , z_j^{\prime})]\cdot\cdot\cdot w(z_n , z_n^{\prime})\\ & \\ = & \sum_{j=1,j\neq i}^n \frac{t_j^a}{z_i - z_j}w(z_1 , z_1^{\prime})\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \end{array } \label{norm3}\ ] ] where for the second equality we have used the definition of tensor product . from ( [ norm3 ] ) and by applying ( [ k9a ] ) to @xmath216 for @xmath229 we have the following knizhnik - zamolodchikov equation @xcite @xcite@xcite : @xmath230 for @xmath231 .
we remark that in ( [ n9 ] ) we have defined @xmath232 and @xmath233\cdot\cdot\cdot [ t^aw(z_j , z_j^{\prime})]\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \end{array } \label{n9a}\ ] ] it is interesting and important that we also have another kz equation with respect to the @xmath234 variables .
the derivation of this kz equation is dual to the above derivation in that the operator products and their corresponding variables are with reverse order to that in the above derivation . from ( [ k2 ] ) and ( [ n8b1 ] )
we have a @xmath235 operator product given by @xmath236 where we have omitted the regular term of the product . then similar to the above derivation of the kz equation from ( [ d2 ] )
we can then derive the following knizhnik - zamolodchikov equation which is dual to ( [ n9 ] ) : @xmath237 for @xmath231 where we have defined : @xmath238\cdot\cdot\cdot [ w(z_j , z_j^{\prime})t^a]\cdot\cdot\cdot w(z_n , z_n^{\prime } ) \end{array } \label{d8a}\ ] ] * rematk*. from the generator @xmath11 and the kac - moody algebra we have derived a quantum kz equation in dual form .
this quantum kz equation in dual form may be consider as a quantum yang - mills equation since it is analogous to the classical yang - mills equation which is derived from the classical yang - mills gauge model .
this quantum kz equation in dual form will be as the starting point for the construction of quantum knots and links . @xmath72
let us consider the following product of two quantum wilson lines : @xmath239 where the two quantum wilson lines @xmath240 and @xmath241 represent two pieces of curves starting at @xmath70 and @xmath242 and ending at @xmath243 and @xmath244 respectively .
we have that this product @xmath245 satisfies the kz equation for the variables @xmath70 , @xmath242 and satisfies the dual kz equation for the variables @xmath243 and @xmath244 .
then by solving the two - variables - kz equation in ( [ n9 ] ) we have that a form of @xmath245 is given by @xcite@xcite@xcite : @xmath246}c_1 \label{m2}\ ] ] where @xmath247 and @xmath248 denotes a constant matrix which is independent of the variable @xmath249 .
we see that @xmath245 is a multivalued analytic function where the determination of the @xmath250 sign depended on the choice of the branch .
similarly by solving the dual two - variable - kz equation in ( [ d8 ] ) we have that @xmath245 is of the form @xmath251 } \label{m3}\ ] ] where @xmath252 denotes a constant matrix which is independent of the variable @xmath253 . from ( [ m2 ] ) , ( [ m3 ] )
and we let @xmath254}$ ] , @xmath255}a$ ] where @xmath256 is a constant matrix we have that @xmath245 is given by @xmath257}ae^{t\log [ \pm ( z_4-z_2 ) ] } \label{m4}\ ] ] where at the singular case that @xmath258 we simply define @xmath259=0 $ ] .
similarly for @xmath260 .
let us find a form of the initial operator @xmath256 .
we notice that there are two operators @xmath261}$ ] and @xmath262 acting on the two sides of @xmath256 respectively where the two independent variables @xmath263 of @xmath264 are mixedly from the two quantum wilson lines @xmath240 and @xmath241 respectively and the the two independent variables @xmath265 of @xmath266 are mixedly from the two quantum wilson lines @xmath240 and @xmath241 respectively . from this
we determine the form of @xmath256 as follows .
let @xmath267 denote a representation of @xmath4 .
let @xmath268 represent an element @xmath49 of @xmath4 and let @xmath269 denote the tensor product representation of @xmath4 .
then in the kz equation we define @xmath270 [ d(g_1)\otimes d(g_1)]\otimes [ d(g_2)\otimes d(g_2 ) ] : = [ t^ad(g_1)\otimes d(g_1)]\otimes [ t^ad(g_2)\otimes d(g_2 ) ] \label{tensorproduct}\ ] ] and @xmath271\otimes [ d(g_2)\otimes d(g_2)][t^a\otimes t^a ] : = [ d(g_1)\otimes d(g_1)t^a]\otimes [ d(g_2)\otimes d(g_2)t^a ] \label{tensorproduct2}\ ] ] then we let @xmath272 denote the universal enveloping algebra where @xmath273 denotes an algebra which is formed by the lie algebra @xmath274 and the identity matrix . now let the initial operator @xmath256 be of the form @xmath275 with @xmath276 taking values in @xmath272 . in this case we have that in ( [ m4 ] ) the operator @xmath261}$ ] acts on @xmath256 from the left via the following formula : @xmath277\otimes a_2\otimes [ t^a a_3]\otimes a_4 \label{ini2}\ ] ] similarly the operator @xmath278}$ ] in ( [ m4 ] ) acts on @xmath256 from the right via the following formula : @xmath279\otimes a_3\otimes[a_4 t^a ] \label{ini3}\ ] ] we may generalize the above tensor product of two quantum wilson lines as follows .
let us consider a tensor product of @xmath0 quantum wilson lines : @xmath280 where the variables @xmath216 , @xmath234 are all independent . by solving the two kz equations we have that this tensor product is given by : @xmath281 where @xmath282 denotes a product of @xmath283 or @xmath262 for @xmath218 where @xmath284 . in ( [ tensor ] )
the initial operator @xmath256 is represented as a tensor product of operators @xmath285 where each @xmath286 is of the form of the initial operator @xmath256 in the above tensor product of two - wilson - lines case and is acted by @xmath283 or @xmath262 on its two sides respectively .
let us consider the following product of two quantum wilson lines : @xmath287 where the two quantum wilson lines @xmath240 and @xmath241 represent two pieces of curves starting at @xmath70 and @xmath242 and ending at @xmath243 and @xmath244 respectively . as shown in the above section we have that @xmath245 is given by the following formula : @xmath257}ae^{t\log [ \pm ( z_4-z_2 ) ] } \label{m4a}\ ] ] where the product is a 4-tensor .
let us set @xmath288 .
then the 4-tensor @xmath289 is reduced to the 2-tensor @xmath290 . by using ( [ m4a ] ) the 2-tensor @xmath290 is given by : @xmath291}a_{14}e^{t\log [ \pm ( z_4-z_2 ) ] } \label{closed1}\ ] ] where @xmath292 is a 2-tensor reduced from the 4-tensor @xmath293 in ( [ m4a ] ) . in this reduction the @xmath294 operator of @xmath295}$ ] acting on the left side of @xmath296 and @xmath297 in @xmath256
is reduced to acting on the left side of @xmath296 and @xmath298 in @xmath299 .
similarly the @xmath294 operator of @xmath300}$ ] acting on the right side of @xmath301 and @xmath298 in @xmath256 is reduced to acting on the right side of @xmath296 and @xmath298 in @xmath299 .
then since @xmath294 is a 2-tensor operator we have that @xmath294 is as a matrix acting on the two sides of the 2-tensor @xmath299 which is also as a matrix with the same dimension as @xmath294 .
thus @xmath302 and @xmath303 are as matrices of the same dimension as the matrix @xmath299 acting on @xmath299 by the usual matrix operation . then since @xmath294 is a casimir operator for the 2-tensor group representation of @xmath4 we have that @xmath304 and @xmath305 commute with @xmath299 since @xmath304 and @xmath303 are exponentials of @xmath294 ( we remark that @xmath304 and @xmath305 are in general not commute with the 4-tensor initial operator @xmath256 ) .
thus we have @xmath306}a_{14}e^{t\log[\pm ( z_4-z_2 ) ] } = e^{-t\log [ \pm ( z_1-z_2)]}e^{t\log[\pm ( z_4-z_2)]}a_{14 } \label{closed1a}\ ] ] we let @xmath290 be as a representation of the quantum wilson line @xmath307 and we write @xmath308
. then we have the following representation of @xmath307 : @xmath309}e^{t\log[\pm ( z_4-w_1)]}a_{14 } \label{closed1a1}\ ] ] this representation of the quantum wilson line @xmath307 means that the line ( or path ) with end points @xmath70 and @xmath244 is specified that it passes the intermediate point @xmath310 .
this representation shows the quantum nature that the path is not specified at other intermediate points except the intermediate point @xmath310 .
this unspecification of the path is of the same quantum nature of the feymann path description of quantum mechanics .
then let us consider another representation of the quantum wilson line @xmath307 .
we consider @xmath311 which is obtained from the tensor @xmath312 by two reductions where @xmath223 , @xmath313 , @xmath314 , @xmath315 are independent variables .
for this representation we have : @xmath316}e^{-t\log [ \pm ( z_1-w_2 ) ] } e^{t\log[\pm ( z_4-w_1)]}e^{t\log[\pm ( z_4-w_2)]}a_{14 } \label{closed1a2}\ ] ] this representation of the quantum wilson line @xmath307 means that the line ( or path ) with end points @xmath70 and @xmath244 is specified that it passes the intermediate points @xmath317 and @xmath318 .
this representation shows the quantum nature that the path is not specified at other intermediate points except the intermediate points @xmath317 and @xmath318 .
this unspecification of the path is of the same quantum nature of the feymann path description of quantum mechanics .
similarly we may represent the quantum wilson line @xmath307 by path with end points @xmath70 and @xmath244 and is specified only to pass at finitely many intermediate points
. then we let the quantum wilson line @xmath307 as an equivalent class of all these representations .
thus we may write @xmath319 .
* remark*. since @xmath299 is a 2-tensor we have that a natural group representation for the wilson line @xmath307 is the 2-tensor group representation of the group @xmath4 .
consider again the product @xmath320 .
we have that @xmath245 is a multivalued analytic function where the determination of the @xmath250 sign depended on the choice of the branch .
let the two pieces of curves be crossing at @xmath321 .
then we have @xmath322 and @xmath323 .
thus we have @xmath324 if we interchange @xmath70 and @xmath242 , then from ( [ h2 ] ) we have the following ordering : @xmath325 now let us choose a branch .
suppose that these two curves are cut from a knot and that following the orientation of a knot the curve represented by @xmath2 is before the curve represented by @xmath3 .
then we fix a branch such that the product in ( [ m4a ] ) is with two positive signs : @xmath326 then if we interchange @xmath70 and @xmath242 we have @xmath327}ae^{t\log(z_4-z_2 ) } \label{h5}\ ] ] from ( [ h4 ] ) and ( [ h5 ] ) as a choice of branch we have @xmath328 where @xmath329 is the monodromy of the kz equation . in ( [ m7a ] )
@xmath70 and @xmath242 denote two points on a closed curve such that along the direction of the curve the point @xmath70 is before the point @xmath242 and in this case we choose a branch such that the angle of @xmath330 minus the angle of @xmath249 is equal to @xmath331 . *
remark*. we may use other representations of @xmath332 .
for example we may use the following representation : @xmath333 then the interchange of @xmath70 and @xmath242 changes only @xmath249 to @xmath330 . thus the formula ( [ m7a ] ) holds .
similarly all other representations of @xmath332 will give the same result .
@xmath72 now from ( [ m7a ] ) we can take a convention that the ordering ( [ h3 ] ) represents that the curve represented by @xmath2 is upcrossing the curve represented by @xmath3 while ( [ h2 ] ) represents zero crossing of these two curves . similarly from the dual kz equation as a choice of branch which is consistent with the above formula we have @xmath334 where @xmath243 is before @xmath244 .
we take a convention that the ordering ( [ m8a ] ) represents that the curve represented by @xmath2 is undercrossing the curve represented by @xmath3 . here along the orientation of a closed curve the piece of curve represented by @xmath2 is before the piece of curve represented by @xmath3 . in this case since the angle of @xmath330 minus the angle of @xmath249 is equal to @xmath331 we have that the angle of @xmath253 minus the angle of @xmath335 is also equal to @xmath331 and this gives the @xmath336 in this formula ( [ m8a ] ) . from ( [ m7a ] ) and ( [ m8a ] ) we have @xmath337 where @xmath70 and @xmath243 denote the end points of a curve which is before a curve with end points @xmath242 and @xmath244 . from ( [ m9 ] )
we see that the algebraic structure of these quantum wilson lines @xmath1 is analogous to the quasi - triangular quantum group @xcite@xcite .
in this section let us apply the above result which is from the kz equation in dual form to derive the skein relation for the homfly polynomial . from this relation
we then have the skein relation for the jones polynomial which is a special case of the homfly polynomial @xcite@xcite@xcite .
it is well known that from the one - side kz equation we can derive a braid group representation which is related to the the derivation of the skein relation for the jones polynomial @xcite@xcite@xcite .
we shall see here that by applying the two kz equations of the kz equation in dual form we also have a way to derive the skein relation for the homfly polynomial .
let us first consider the following theorem of kohno and drinfield @xcite@xcite@xcite : let @xmath7 be the monodromy of the kz equation for the group @xmath4 and let @xmath338 denotes the @xmath7-matrix of the quantum group @xmath339 where @xmath274 denotes the lie algebra of @xmath4 and @xmath340 where @xmath190 .
then there exists a twisting @xmath341 such that @xmath342 from this relation we have that the braid group representations obtained from the quantum group @xmath343 and obtained from the one - side kz equation are equivalent .
we shall use only the relation ( [ r ] ) of this theorem to derive the skein relation of the homfly polynomial . from the property of the quantum group @xmath344 we also have the following formula @xcite@xcite@xcite :
@xmath345 thus we have @xmath346 by using this formula we have @xmath347 fw(z_1,w)w(w , z_2)w(z_3,w)w(w , z_4)f^{-1 } = 0 \label{h6b}\ ] ] thus by using the relation ( [ r ] ) we have @xmath348 then by using the formulas ( [ m7a ] ) and ( [ m8a ] ) for upcrossing and undercrossing from ( [ h6a ] ) we have @xmath349 let us make a further twist that replace @xmath350 by @xmath351 where @xmath352 denotes a nonzero variable
. then from ( [ h7 ] ) we have the following skein relation for the homfly polynomial : @xmath353 where we define @xmath354 and that @xmath355 , @xmath356 and @xmath357 are defined by @xmath358 which are as the homfly polynomials for upcrossing , zero crossing and undercrossing respectively .
let us consider again the quantum wilson line @xmath308 .
let us set @xmath359 .
in this case the quantum wilson line forms a closed loop . now in ( [ closed1a ] ) with @xmath359 we have that @xmath360 and @xmath361 which come from the two - side kz equations cancel each other and from the multivalued property of the @xmath362 function we have @xmath363 where @xmath329 is the monodromy of the kz equation @xcite .
* remark*. it is clear that if we use other representation of the quantum wilson loop @xmath364 ( such as the representation @xmath365 ) then we will get the same result as ( [ closed2 ] ) .
* remark*. for simplicity we shall drop the subscript of @xmath299 in ( [ closed2 ] ) and simply write @xmath366 .
now we have that the quantum wilson loop @xmath367 corresponds to a closed curve in the complex plane with starting and ending point @xmath70 .
let this quantum wilson loop @xmath367 represents the unknot .
we shall call @xmath367 as the quantum unknot .
then from ( [ closed2 ] ) we have the following invariant for the unknot : @xmath368 where @xmath369 is a 2-tensor constant matrix operator . in the following
let us extend the definition ( [ m6 ] ) to a knot invariant for nontrivial knots .
let @xmath370 represent a piece of curve with starting point @xmath216 and ending point @xmath223 .
then we let @xmath371 represent two pieces of uncrossing curve .
then by interchanging @xmath70 and @xmath242 we have @xmath372 represent the curve specified by @xmath2 upcrossing the curve specified by @xmath3 .
now for a given knot diagram we may cut it into a sum of parts which are formed by two pieces of curves crossing each other . each of these parts is represented by ( [ m12 ] ) ( for a knot diagram of the unknot with zero crossings we simply do not need to cut the knot diagram ) .
then we define the trace of a knot with a given knot diagram by the following form : @xmath373 where we use ( [ m12 ] ) to represent the state of the two pieces of curves specified by @xmath2 and @xmath3 .
the @xmath374 means the product of a sequence of parts represented by ( [ m12 ] ) according to the state of each part .
the ordering of the sequence in ( [ m14 ] ) follows the ordering of the parts given by the orientation of the knot diagram .
we shall call the sequence of crossings in the trace ( [ m14 ] ) as the generalized wilson loop of the knot diagram . for the knot diagram of the unknot with zero crossings we simply let it be @xmath6 and call it the quantum wilson loop .
we shall show that the generalized wilson loop of a knot diagram has all the properties of the knot diagram and that ( [ m14 ] ) is a knot invariant . from this we shall call a generalized wilson loop as a quantum knot .
before the proof that a generalized wilson loop of a knot diagram has all the properties of the knot diagram in the following let us first consider some examples to illustrate the way to define ( [ m14 ] ) and the way of applying the braiding formulas ( [ m7a ] ) , ( [ m8a ] ) and ( [ m9 ] ) to equivalently transform ( [ m14 ] ) to a simple expression of the form @xmath375 where @xmath8 is an integer .
let us first consider the knot in fig.1 .
for this knot we have that ( [ m14 ] ) is given by @xmath376 where the product of quantum wilson lines is from the definition ( [ m12 ] ) represented a crossing at @xmath321 . in applying ( [ m12 ] )
we let @xmath70 be the starting and the ending point .
fig.1 then we have that ( [ m15a ] ) is equal to @xmath377 where we have used ( [ m9 ] ) .
we see that ( [ m16 ] ) is just the knot invariant ( [ m6 ] ) of the unknot .
thus the knot in fig.1 is with the same knot invariant of the unknot and this agrees with the fact that this knot is topologically equivalent to the unknot .
let us then consider a trefoil knot in fig.2a . by ( [ m12 ] ) and
similar to the above examples we have that the definition ( [ m14 ] ) for this knot is given by : @xmath378 where we have repeatly used ( [ m9 ] )
. then we have that ( [ m21 ] ) is equal to : @xmath379 where we have used ( [ m7a ] ) and ( [ m9 ] ) .
this is as a knot invariant for the trefoil knot in fig.2a .
fig.2a fig.2b then let us consider the trefoil knot in fig . 2b which is the mirror image of the trefoil knot in fig.2a .
the definition ( [ m14 ] ) for this knot is given by : @xmath380 where similar to ( [ m21 ] ) we have repeatly used ( [ m9 ] ) .
then we have that ( [ m25 ] ) is equal to : @xmath381 where we have used ( [ m8a ] ) and ( [ m9 ] ) .
this is as a knot invariant for the trefoil knot in fig.2b .
we notice that the knot invariants for the two trefoil knots are different .
this shows that these two trefoil knots are not topologically equivalent .
more examples of the above quantum knots and knot invariants will be given in a following section .
let us now show that the generalized wilson loop of a knot diagram has all the properties of the knot diagram and that ( [ m14 ] ) is a knot invariant . to this end
let us first consider the structure of a knot .
let @xmath382 be a knot .
then a knot diagram of @xmath382 consists of a sequence of crossings of two pieces of curves cut from the knot @xmath382 where the ordering of the crossings can be determined by the orientation of the knot @xmath382 . as an example we may consider the two trefoil knots in the above section .
each trefoil knot is represented by three crossings of two pieces of curves .
these three crossings are ordered by the orientation of the trefoil knot starting at @xmath70 .
let us denote these three crossings by @xmath383 , @xmath384 and @xmath385 .
then the sequence of these three crossings is given by @xmath386 .
on the other hand if the ordering of the three crossings starts from other @xmath216 on the knot diagram then we have sequences @xmath387 and @xmath388 .
all these sequences give the same knot diagram and they can be transformed to each other by circling as follows : @xmath389 where @xmath390 means that the number @xmath352 is to be moved to the @xmath390 position as indicated .
let us call ( [ s ] ) as the circling property of the trefoil knot . as one more example let us consider the figure - eight knot in fig.3 .
the simplest knot diagram of this knot has four crossings .
fig.3 starting at @xmath70 let us denote these crossings by 1 , 2 , 3 and 4 .
then we have the following circling property of the figure - eight knot : @xmath391 we notice that in this cirling of the figure - eight knot there are subcirclings .
in summary we have that a knot diagram of a knot @xmath382 can be characterized as a finite sequence of crossings of curves which are cut from the knot diagram where the ordering of the crossings is derived from the orientation of the knot diagram and has a circling property for which ( [ s ] ) and ( [ s2 ] ) are examples .
now let us represent a knot diagram of a knot @xmath382 by a sequence of products of wilson lines representing crossings as in the above section .
let us call these products of wilson lines by the term w - product .
then we call this sequence of @xmath392-products as the generalized wilson loop of the knot diagram of a knot @xmath382 .
let us consider the following two @xmath392-products : @xmath393 in the above section we have shown that these two w - products faithfully represent two oriented pieces of curves crossing or not crossing each other where @xmath2 and @xmath3 represent these two pieces of curves .
now there is a natural ordering of the @xmath392-products of crossings derived from the orientation of a knot as follows .
let @xmath2 and @xmath3 represent two pieces of curves where the piece of curve represented by @xmath2 is before the piece of curve represented by @xmath3 according to the orientation of a knot .
then the ordering of these two pieces of curves can be represented by the product @xmath332 .
now let 1 and 2 denote two @xmath392-products of crossings where we let 1 before 2 according to the orientation of a knot .
then from the ordering of pieces of curves we have that the product @xmath394 represents the ordering of the two crossings 1 and 2 .
now let a knot diagram of a knot @xmath382 be given .
let the crossings of this knot diagram be denoted by @xmath395 and let this knot diagram be characterized by the sequence of crossings @xmath396 which is formed according to the orientation of this knot diagram .
on the other hand let us for simplicity also denote the corresponding @xmath392-products of crossings by @xmath395 .
then the whole product of @xmath392-products of crossings @xmath396 represents the sequence @xmath396 of crossings which is identified with the the knot diagram .
this whole product @xmath397 of @xmath392-products of crossings is the generalized wilson loop of the knot diagram and we denote it by @xmath398 . in
the following let us show that this generalized wilson loop @xmath398 has the circling property of the sequences of crossings of the knot diagram .
it then follows that this generalized wilson loop represents all the properties of the sequence @xmath396 of crossings of the knot diagram .
then since this sequence @xmath396 of crossings of the knot diagram is identified with the knot diagram we have that this generalized wilson loop @xmath398 can be identified with the knot diagram and we have the following theorem . each knot @xmath382
can be faithfully represented by its generalized wilson loop @xmath398 in the sense that if two knot diagrams have the same generalized wilson loop then these two knot diagrams must be topologically equivalent
. * proof .
* let us show that the generalized wilson loop @xmath398 of a knot diagram of @xmath382 has the circling property .
let us consider a product @xmath399 where we first let @xmath400 and @xmath244 be all independent . by solving the two kz equations as shown in the above sections we have @xmath401 } ae^{t \log [ \pm ( z_2-z_4 ) ] } \label{sa}\ ] ] where the initial operator @xmath256 is a 4-tensor as shown in the above sections .
the sign @xmath250 in ( [ sa ] ) reflects that solutions of the kz equations are complex multi - valued functions .
( we remark that the 4-tensor initial operator @xmath256 in general may not commute with @xmath402}$ ] and @xmath403}$ ] ) .
then the interchange of @xmath404 and @xmath3 corresponds to that @xmath70 and @xmath242 interchange their positions and @xmath243 and @xmath244 interchange their positions respectively .
this interchange gives a pair of sign changes : @xmath405 from this we have that @xmath406 is given by @xmath407 } ae^{t \log [ \pm ( z_4-z_2 ) ] } \label{sb}\ ] ] now let us set @xmath288 and @xmath359 such that the two products in ( [ sa ] ) and ( [ sb ] ) form a closed loop .
in this case we have that the initial operator @xmath256 is reduced from a 4-tensor to a 2-tensor and that @xmath264 and @xmath266 act on @xmath256 by the usual matrix operation where @xmath256 , @xmath264 and @xmath266 are matrices of the same dimension .
in this case we have that @xmath256 commutes with @xmath264 and @xmath266 since @xmath264 and @xmath266 are casimir operators on @xmath4 .
let us take a definite choice of branch such that the sign change @xmath408 gives a @xmath409 difference from the multivalued function @xmath362 .
then we have that @xmath410 . then since @xmath240 and @xmath241 represent two lines with @xmath70 , @xmath243 and @xmath242 , @xmath244 as starting and ending points respectively we have that the sign change @xmath411 also gives the same @xmath409 difference from the multivalued function @xmath362 .
thus we have that @xmath412 .
it follows from this pair of sign changes and that @xmath256 commutes with @xmath264 and @xmath266 we have that @xmath413 when @xmath288 and @xmath359 .
this proves the simplest circling property of generalized wilson loops .
we remark that in the above proof the pair of sign changes gives two factors @xmath7 and @xmath336 which cancel each other and gives the circling property .
we shall later apply the same reason of pair sign changes to get the general circling property .
we also remark that the proof of this circling property is based on the same reason as the derivation of the braiding formulas ( [ m7a ] ) , ( [ m8a ] ) and ( [ m9 ] ) as shown in the above sections .
let us consider a product of @xmath0 quantum wilson lines @xmath414 , @xmath229 , with the property that the end points @xmath216 , @xmath234 of these quantum wilson lines are connected to form a closed loop . from the analysis in the above sections we have that this product is reduced from a tensor product to a 2-tensor .
it then follows from ( [ tensor ] ) that this product is of the following form : @xmath415 where the initial operator @xmath256 is reduced to a 2-tensor and that the @xmath250 signs of @xmath283 and @xmath416 are to be determined .
then since @xmath283 and @xmath416 commute with @xmath256 we can write ( [ class ] ) in the form @xmath417 where @xmath284 . from this formula let us derive the general circling property as follows .
let us consider two generalized wilson lines denoted by 1 and 2 respectively . here by the term generalized wilson line we mean a product of quantum wilson lines with two open ends .
as a simple example let us consider the product @xmath418 . by definition
this is a generalized wilson line with two open ends @xmath70 and @xmath243 ( @xmath38 is not an open end ) .
suppose that the two open ends of 1 and 2 are connected .
then we want to show that @xmath419 .
this identity is a generalization of the above interchange of @xmath2 and @xmath3 with @xmath288 and @xmath359 . because @xmath394 and @xmath420 form closed loops we have that @xmath394 and @xmath420 are products of quantum wilson lines @xmath421 ( where @xmath422 and @xmath423 denote some @xmath424 or @xmath425 where we use @xmath425 to denote crossing points ) such that for each pair of variables @xmath422 and @xmath314 appearing at the left side of @xmath421 and @xmath426 there is exactly one pair of variables @xmath422 and @xmath314 appearing at the right side of @xmath427 and @xmath428 .
thus in the formula ( [ basic ] ) ( with the variables @xmath38 , @xmath103 in ( [ basic ] ) denoted by variables @xmath429 ) we have that the factors @xmath430 and @xmath431 appear in pairs . as in the above case we have that the interchange of the open ends of @xmath394 and @xmath420 interchanges @xmath394 to @xmath420 .
this interchange gives changes of the factors @xmath430 and @xmath431 as follows .
let @xmath70 and @xmath243 be the open ends of @xmath383 and @xmath242 and @xmath244 be the open ends of @xmath384 such that @xmath359 and @xmath288 . consider a factor @xmath432 .
the interchange of @xmath70 and @xmath242 interchanges this factor to @xmath433 .
then there is another factor @xmath434 .
the interchange of @xmath243 and @xmath244 interchanges this factor to @xmath435 .
thus this is a pair of sign changes . by the same reason and the consistent choice of branch as in the above case we have that the formula ( [ basic ] ) is unchanged under this pair of sign changes . then let us consider a factor @xmath430 of the form @xmath436 where @xmath437 and @xmath314 is not an open end .
corresponding to this factor we have the factor @xmath438 .
then under the interchange of @xmath70 and @xmath242 we have that @xmath436 and @xmath438 change to @xmath438 and @xmath436 respectively which gives no change to the formula ( [ basic ] ) .
a similar result holds for the interchange of @xmath243 and @xmath244 for factor @xmath439 and @xmath440 .
it follows that under the interchange of the open ends of @xmath383 and @xmath384 we have the pairs of sign changes from which the formula ( [ basic ] ) is unchanged .
this shows that @xmath441 .
then we consider two generalized wilson products of crossings which are products of crossings with four open ends respectively .
let us again denote them by @xmath383 and @xmath384 .
each such generalized wilson crossing can be regarded as the crossing of two generalized wilson lines .
then the interchange of two open ends of the two generalized lines of @xmath383 with the two open ends of the two generalized lines of @xmath384 respectively interchanges @xmath394 to @xmath420 .
then let us suppose that the open ends of these two wilson products are connected in such a way that the products @xmath394 and @xmath420 form closed loops .
in this case we want to show that @xmath419 which is a circling property of a knot diagram .
the proof of this equality is again similar to the above cases . in this case
we also have that the interchange of the open ends of the two generalized wilson crossings gives pairs of sign changes of the factors @xmath430 and @xmath431 in @xmath442 and @xmath442 .
then by using ( [ basic ] ) we have @xmath443 .
let us then consider two generalized wilson products of crossings denoted by @xmath383 and @xmath384 with open ends connected in such a way that two open ends of @xmath383 are connected to two open ends of @xmath384 to form a closed loop .
we want to prove that @xmath419 .
this will give the subcircling property .
since a closed loop is formed we have that each open end of @xmath383 or of @xmath384 is connected to a closed loop . in this case as the above cases we have that the products @xmath394 is with the initial operator @xmath256 being a 2-tensor since the open ends of @xmath383 or @xmath384 do not cause @xmath256 to be a tensor with tensor degree more than @xmath384 by their connection to the closed loop .
indeed , let @xmath38 be an open end of @xmath383 or @xmath384 .
then it is an end point of a quantum wilson line @xmath1 which is a part of @xmath383 and @xmath384 such that @xmath103 is on the closed loop formed by @xmath383 and @xmath384 .
then we have that this quantum wilson line @xmath1 is connected with the closed loop at @xmath103 .
since the loop is closed we have that this quantum wilson line @xmath1 and the closed loop are connected into a connected line with orientation .
it follows that the open end @xmath38 gives no additional tensor degree to the initial operator @xmath256 for the product @xmath394 or @xmath420 . and that the initial operator @xmath256 is still as the initial operator for the closed loop that it is a 2-tensor ( we remark that in the above section on computation of quantum wilson loop we see that an open quantum wilson line @xmath307 and a closed quantum wilson loop @xmath364 are with the same 2-tensor initial operator @xmath256 .
this shows that the open end @xmath70 of a quantum wilson line @xmath307 gives no additional tensor degree to the initial operator @xmath256 of the closed quantum wilson loop @xmath364 .
this is the same reason that the open end @xmath38 of the quantum wilson line @xmath1 gives no additional tensor degree to the initial operator @xmath256 for the product @xmath394 or @xmath420 ) .
now since @xmath256 is a 2-tensor we have that @xmath256 , @xmath264 and @xmath266 are as matrices of the same dimension .
in this case we have that @xmath256 commutes with @xmath264 and @xmath266 . then by interchange the open ends of @xmath383 with open ends of @xmath384 we interchange @xmath394 to @xmath420 .
this interchange again gives pairs of sign changes .
then since the initial operator @xmath256 commutes with @xmath264 and @xmath266 we have that @xmath419 , as was to be proved
. then we let @xmath394 and @xmath420 be connected to another generalized wilson product of crossing denoted by @xmath385 to form a closed loop .
then from @xmath419 we have @xmath444 and @xmath445 .
this gives the subcircling property of generalized wilson loops .
this subcircling property has been illustrated in the knot diagram of the fight - eight knot .
then from a case in the above we also have the circling property @xmath446 between @xmath385 and @xmath420 . continuing in this way we have the circling or subcircling properties for generalized wilson loops whenever
the open ends of a product of generalized wilson lines or crossings are connected in such a way that among the open ends a closed loop is formed .
this shows that the generalized wilson loop of a knot diagram has the circling property of the knot diagram . with this circling property
it then follows that the generalized wilson loop of a knot diagram completely describes the structure of the knot diagram .
now since the generalized wilson loop of a knot diagram is a complete copy of this knot diagram we have that two knot diagrams which can be equivalently moved to each other if and only if the corresponding generalized wilson loops can be equivalently moved to each other .
thus we have that if two knot diagrams have the same generalized wilson loop then these two knot diagrams must be equivalent .
this proves the theorem .
@xmath72 * examples of generalized wilson loops * as an example of generalized wilson loops let us consider the trefoil knots .
starting at @xmath70 let the w - product of crossings be denoted by 1 , 2 and 3 .
then we have the following circling property of the generalized wilson loops of the trefoil knots : @xmath447 as one more example let us consider the figure - eight knot .
starting at @xmath70 let the w - product of crossings be denoted by 1 , 2 , 3 and 4 .
then we have the following circling property of the generalized wilson loop of the figure - eight knot : @xmath448 @xmath72 * definition * we may call a generalized wilson loop of a knot diagram as a quantum knot since by the above theorem this generalized wilson loop is a complete copy of the knot diagram .
@xmath72 from the above theorem we have the following theorem .
let @xmath398 denote the generalized wilson loop of a knot @xmath382 .
then we can write @xmath398 in the form @xmath449 for some integer @xmath8 where @xmath450 denotes a trivial knot and @xmath451 denotes a wilson loop on @xmath450 with starting point @xmath70 and ending point @xmath70 . from this form
we have that the trace @xmath452 is a knot invariant which classifies knots .
thus knots can be classified by the integer @xmath8 .
* proof . *
since a generalized wilson loop @xmath453 is in a closed and connected form we have that a generalized wilson loop @xmath453 can be of the form ( [ basic ] ) .
thus from the multivalued property of the @xmath362 function and the two - side cancelation in ( [ basic ] ) we have that @xmath453 can be of the following ( multivalued ) form @xmath454 for some integer @xmath455 .
furthermore for nontrivial knot @xmath382 there are some factors @xmath456 of @xmath457 coming from the braidings of wilson lines ( for which the generalized wilson loop @xmath398 is formed ) by braiding operations such as ( [ m7a])and ( [ m8a ] ) .
thus we can write the integer @xmath162 in the form @xmath458 for some integer @xmath8 and for some integer @xmath459 where @xmath0 is obtained by the two - side cancelations in such a way that the cancelations are obtained when the wilson lines of the knot diagram for @xmath382 are connected together to form a wilson loop @xmath460 where @xmath450 is a closed curve which is as an unknot and is of the same form as the knot diagram for @xmath382 when this knot diagram of @xmath382 is considered only as a closed curve in the plane ( such that the upcrossings and undercrossings are changed to let @xmath382 be the unknot @xmath450 ) . from this
we have @xmath461 for @xmath462 .
thus @xmath398 can be written in the following form for some @xmath8 : @xmath463 this number @xmath8 is unique since if there is another number @xmath464 such that @xmath465 then we have the equality : @xmath466 this shows that @xmath467 and thus @xmath468 . from ( [ i ] )
we also have @xmath469 for some integer @xmath8 and that @xmath470 is a knot invariant .
then let us show that the invariant @xmath470 classifies knots .
let @xmath471 and @xmath472 be two knots with the same invariant @xmath470
. then @xmath471 and @xmath472 are both with the same invariant @xmath473 where the trace is omitted . then by the above formula ( [ i ] ) we have @xmath474 thus @xmath475 and @xmath476 can be transformed to each other .
thus @xmath471 and @xmath472 are equivalent .
thus the invariant @xmath477 classifies knots .
it follows that the invariant @xmath452 classifies knots and thus knots can be classified by the integer @xmath8 , as was to be proved .
in this section let us give more computations of the knot invariant @xmath478
. we shall show by computation ( with the chosen braiding formulae ) that the fight - eight knot @xmath479 is assigned with the number @xmath480 and two composite knots composed by two trefoil knots ( with the names reef knot and granny knot and denoted by @xmath481 and @xmath482 respectively ) are assigned with the numbers @xmath483 and @xmath484 respectively .
the computation is quite tedious . in the next section
we shall have a more efficient way to determine the integer @xmath8 .
readers may skip this section for the first reading .
let us first consider the figure - eight knot . from the figure of this knot in a above section we have that the knot invariant of this knot is given by : @xmath485 in the above computation we have chosen @xmath70 as the staring point ( by the circling property we may choose any point as the starting point ) . by repeatedly applying the braiding formulas ( [ m7a]),([m8a ] ) and ( [ m9 ] ) we have that this invariant is equal to : @xmath486 then we have that ( [ more2 ] ) is equal to @xmath487 where @xmath488 with @xmath38 being a point on the line represented by @xmath489 and that @xmath490 . since @xmath70 is as the starting and ending point and
is an intermediate point we have the following braiding formula : @xmath491 thus we have that ( [ more2 ] ) is equal to @xmath492 then in ( [ more5 ] ) we have that @xmath493 this shows that @xmath494 is a generalized wilson line . then since generalized wilson lines are with the same braiding formulas as wilson lines we have that by a braiding formula similar to ( [ more4 ] ) ( for @xmath70 as the starting and ending point and as an intermediate point ) the formula ( [ more5 ] ) is equal to : @xmath495 where the first equality is by a braiding formula which is similar to the braiding formula ( [ more4 ] ) .
this is the knot invariant for the figure - eight knot and we have that @xmath480 for this knot .
let us then consider the composite knot @xmath481 in fig.4 .
the trace of the generalized loop of this knot is given by ( in fig.4 one of the two @xmath496 should be @xmath497 ) : @xmath498 fig.4 by repeatedly applying braiding formula ( [ m9 ] ) we have that this invariant is equal to @xmath499 where the braiding of @xmath500 and @xmath501 gives @xmath502 .
this braiding formula comes from the fact that the wilson line @xmath501 represents a curve with end points @xmath318 and @xmath503 such that one and a half loop is formed which can not be removed because the end point @xmath503 is attached to this curve itself to form the closed loop .
this closed loop gives a @xmath504 phase angle which is a topological effect .
thus while the usual braiding of two pieces of curves gives @xmath7 which is of a @xmath331 phase angle we have that the braiding of @xmath500 and @xmath501 gives @xmath7 and an additional @xmath504 phase angle and thus gives @xmath502 . then we have that ( [ more8 ] ) is equal to @xmath505 this is the invariant of @xmath481 .
thus we have that @xmath483 for @xmath481 .
let us then consider the composite knot @xmath482 in fig.5 .
we have that the trace of the generalized wilson loop of @xmath482 is given by ( in fig.5 one of the two @xmath496 should be @xmath497 ) : @xmath506 fig.5 by repeatedly applying braiding formulas ( [ m7a ] ) , ( [ m8a ] ) and ( [ m9 ] ) we have that this invariant is equal to @xmath507 where the quantum wilson line @xmath508 represents the piece of curve which starts at @xmath317 and goes through @xmath509 , @xmath510 , @xmath70 and ends at @xmath511 .
this curve includes a one and a half loop which can not be removed since @xmath317 is attached to this curve to form the loop .
this is of the same case as that in the knot @xmath481 .
this is a topological property which gives a @xmath504 phase angle .
we have that ( [ more11 ] ) is equal to @xmath512 where the piece of curve represented by quantum wilson line @xmath513 also contains the closed loop .
now let this knot @xmath482 be starting and ending at @xmath514 . then by the braiding formula on @xmath513 and @xmath515 as in the case of the knot
@xmath481 we have that ( [ more12 ] ) is equal to @xmath516 where we have repeatedly applied the braiding formula ( [ m9 ] ) .
now let @xmath511 be the starting and ending point .
then we have that ( [ more13 ] ) is equal to @xmath517 where @xmath518 denotes the following generalized wilson line : @xmath519 then by the same braiding formula for generalized wilson lines as that for wilson lines ( with @xmath511 as the starting and ending point and as an intermediate point ) we have that ( [ more14 ] ) is equal to : @xmath520 this is the knot invariant for the knot @xmath482 .
thus we have that @xmath484 for the knot @xmath482
. then we have that the image of @xmath482 is with the knot invariant @xmath521 .
in the above sections the computations of the knot invariant @xmath478 is tedious . in this section
let us use another method to determine the integer @xmath8 without carrying out the tedious computations .
we shall use only the connected sum operation on knots to find out the integer @xmath8 . for simplicity
we use the positive integer @xmath522 to form a classification table of knots where @xmath8 is assigned to a knot while @xmath523 is assigned to its mirror image if the knot is not equivalent to its mirror image .
our main references on the connected sum operation on knots are @xcite-@xcite .
let @xmath524 denote the connected sum of two knots such that the resulting total number of alternating crossings is equal to the sum of alternating crossings of each of the two knots minus @xmath384 .
as an example we have the reef knot ( or the square knot ) @xmath525 which is a composite knot composed with the knot @xmath526 and its mirror image as in fig.4 .
this square knot has @xmath527 crossings and @xmath528 alternating crossings .
then let @xmath529 denote the connected sum for two knots such that the resulting total number of alternating crossings is equal to the sum of alternating crossings of each of the two knots .
as an example we have the granny knot @xmath482 which is a composite knot composed with two identical knots @xmath526 as in fig.5 ( for simplicity we use one notation @xmath526 to denote both the trefoil knot and its mirror image though these two knots are nonequivalent ) .
this knot has @xmath527 alternating crossings which is equal to the total number of crossings .
we have that the two operations @xmath524 and @xmath529 satisfy the commutative law and the associative law @xcite-@xcite .
further for each knot there is a unique factorization of this knot into a @xmath524 and @xmath529 operations of prime knots which is similar to the unique factorization of a number into a product of prime numbers @xcite-@xcite .
we shall show that there is a deeper connection between these two factorizations .
we shall show that we can establish a classification table of knots where each knot is assigned with a number such that prime knots are bijectively assigned with prime numbers such that the prime number @xmath384 corresponds to the trefoil knot ( the trefoil knot will be assigned with the number @xmath383 and is related to the prime number @xmath384 ) .
we have shown by computation that the knot @xmath530 is with @xmath531 , the knot @xmath532 is with @xmath480 .
thus there are no knots assigned with the number @xmath384 since other knots are with crossings more than these two knots .
we have shown by computation that the knot @xmath481 is assigned with the number @xmath528 .
thus we have @xmath533 ( since knots are assigned with integers we may regard the @xmath524 and @xmath529 as operations on the set of numbers ) .
this shows that the number @xmath383 plays the role of the number @xmath384 .
thus while the knot @xmath530 is with @xmath531 we may regard this @xmath531 is as the even prime number @xmath384 .
we shall have more to say about this phenomenon of @xmath383 and @xmath384 .
this phenomenon reflects that the operation @xmath524 has partial properties of addition and multiplication where @xmath531 is assigned to @xmath534 for addition while @xmath534 plays the role of @xmath384 is for multiplication .
the aim of this section is to find out a table of the relation between knots and numbers by using only the operations @xmath524 and @xmath529 on knots and by using the following data as the initial step for induction : * initial data for induction : * the prime knot @xmath534 is assigned with the number @xmath383 and it also plays the role of @xmath384 .
this means that the number @xmath384 is not assigned to other knots and is left for the prime knot @xmath534 . @xmath72
* remark*. we shall say that the prime knot @xmath534 is assigned with the number @xmath383 and is related to the prime number @xmath384 .
@xmath72 we shall give an induction on the number @xmath0 of @xmath535 for establishing the table .
for each induction step on @xmath0 because of the special role of the trefoil knot @xmath534 we let the composite knot @xmath536 obtained by repeatedly taking @xmath524 operation @xmath537 times on the trefoil knot @xmath534 be assigned with the number @xmath535 in this induction .
let us first give the following table relating knots and numbers up to @xmath538 as a guide for the induction for establishing the whole classification table of knots : @xmath539 from this table we see that the @xmath524 operation is similar to the usual multiplication @xmath540 on numbers . without the @xmath529 operation
this @xmath524 operation would be exactly the usual multipilcation on numbers if this @xmath524 operation is regarded as an operation on numbers . from this table
we see that comparable composite knots ( in a sense from the table and we shall discuss this point later ) are grouped in each of the intervals between two prime numbers .
it is interesting that in each interval composite numbers are one - to - one assigned to the comparable composite knots while prime numbers are one - to - one assigned to prime knots . here
a main point is to introduce the @xmath529 operation while keeping composite knots correspond to composite numbers and prime knots correspond to prime numbers . to this end
we need to have rooms at the positions of composite numbers for the introduction of composite knots obtained by the @xmath529 operation .
we shall show that these rooms can be obtained by using the special property of the trefoil knot which is assigned with the number @xmath383 ( for the addition property of the @xmath524 and @xmath529 operations ) while this trefoil knot is similar to the number @xmath384 for the multiplication property of the @xmath524 operation .
let us then carry out the induction steps for obtaining the whole table . to this end
let us investigate in more detail the above comparable properties of knots .
we have the following definitions and theorems .
* definition*. we write @xmath541 if @xmath471 is before @xmath472 in the ordering of knots ; i.e. the number assigned to @xmath471 is less than the number assigned to @xmath472 .
* definition ( preordering)*. let two knots be written in the form @xmath542 and @xmath543 where we have determined the ordering of @xmath472 and @xmath544 . then we say that @xmath542 and @xmath543 are in a preordering in the sense that we put the ordering of these two knots to follow the ordering of @xmath472 and @xmath544 .
if this preordering is not changed by conditions from other preorderings on these two knots ( which are from other factorization forms of these two knots ) then this preordering becomes the ordering of these two knots .
we shall see that this preordering gives the comparable property in the above table . @xmath545 * remark*. a ) this definition is consistent since if @xmath471 is the unknot then we have @xmath542=@xmath472 and @xmath543=@xmath544 and thus the ordering of @xmath542 and @xmath543 follows the ordering of @xmath472 and @xmath544 .
\b ) we can also define similarly the preordering of two knots @xmath546 and @xmath547 with the @xmath529 operation .
@xmath72 we have the following theorem .
consider two knots of the form @xmath542 and @xmath543 where @xmath471 , @xmath472 and @xmath544 are prime knots such that @xmath548 .
then we have @xmath549 . [ pre ] * proof*. since @xmath471 , @xmath472 and @xmath544 are prime knots there are no other factorization forms of the two knots @xmath542 and @xmath543 .
thus these two forms of the two knots are the only way to give preordering to the two knots and thus there are no other conditions to change the preordering given by this factorization form of the two knots .
thus we have that @xmath548 implies @xmath549 .
@xmath72 suppose two knots are written in the form @xmath542 and @xmath543 for determining their ordering and that the other forms of these two knots are not for determining their ordering .
suppose that @xmath548 .
then we have @xmath549 .
* proof*. the proof of this theorem is similar to the proof of the above theorem . since the other factorization forms are not for the determination of the ordering of the two knots in the factorization form @xmath542 and @xmath543 we have that the preordering of these two knots in this factorization form becomes the ordering of these two knots .
thus we have @xmath549 .
@xmath72 as a generalization of theorem [ pre ] we have the following theorem .
let two knots be of the form @xmath542 and @xmath543 where @xmath472 and @xmath544 are prime knots .
suppose that @xmath548 .
then we have @xmath549 . [ pre2 ]
* proof*. we have the preordering that @xmath542 is before @xmath543 . then since @xmath472 and @xmath544 are prime knots we have that the other preordering of @xmath542 and @xmath543 can only from the factorization of @xmath471 . without loss of generality
let us suppose that @xmath471 is of the form @xmath550 where @xmath551 and @xmath552 and @xmath553 are prime knots .
then we have the factorization @xmath554 and @xmath555 .
this factorization is the only factorization that might change the preordering that @xmath542 is before @xmath543 .
then if @xmath556 or @xmath557 with this factorization the two knots @xmath542 and @xmath543 are noncomparable in the sense that this factorization gives no preordering property and that the ordering of these two knots is determined by other conditions .
thus this factorization of @xmath542 and @xmath543 is not for the determination of the ordering of @xmath542 and @xmath543 .
thus the preordering that @xmath542 is before @xmath543 is the ordering of @xmath542 and @xmath543 . on the other hand
if @xmath558 and @xmath559 then this factorization gives the same preordering that @xmath542 is before @xmath543 .
thus for this case the preordering that @xmath542 is before @xmath543 is also the ordering of @xmath542 and @xmath543 .
thus we have @xmath560 .
@xmath72 in addition to the above theorems we have the following theorems .
consider two knots of the form @xmath546 and @xmath547 where @xmath471 , @xmath472 and @xmath544 are prime knots such that @xmath548 .
then we have @xmath561 .
[ 1 ] * proof*. by using a preordering property for knots with @xmath529 operation as similar to that for knots with @xmath524 operation we have that the proof of this theorem is similar to the proof of the above theorems .
@xmath72 let two knots be of the form @xmath546 and @xmath547 where @xmath472 and @xmath544 are prime knots .
suppose that @xmath548 .
then we have @xmath561 .
* proof*. the proof of this theorem is also similar to the proof of the theorem [ pre2 ] .
@xmath72 these two theorems will be used for introducing and ordering knots involved with a @xmath529 operation which will have the effect of pushing out composite knots with the property of jumping over ( to be defined ) such that knots are assigned with a prime number if and only if the knot is a prime knot .
let us investigate more on the property of preordering .
we consider the following * definition ( preordering sequences)*. at the @xmath0th induction step let the prime knot @xmath534 take a @xmath524 operation with the previous @xmath562th step .
we call this obtained sequence of composite knots as a preordering sequence . thus from the ordering of the @xmath562th step we have a sequence of composite knots which will be for the construction of the @xmath0th step
. then we let the prime knot @xmath479 ( or the knot assigned with a prime number which is @xmath385 in the @xmath384nd step as can be seen from the above table ) take a @xmath524 operation with the previous @xmath563th step . from this
we get a sequence of composite knots for constructing the @xmath0th step .
then we let the prime knots @xmath564 and @xmath565 ( which are prime knots in the same step assigned with a prime number which is @xmath566 or @xmath567 in the @xmath385rd step as can be seen from the above table ) take a @xmath524 operation with the previous @xmath568th step respectively . from this
we get two sequences for constructing the @xmath0th step . continuing in this way until the sequences are obtained by a prime knot in the @xmath562th step taking a @xmath524 operation with the step @xmath569 where the prime knot is assigned with a prime number in the @xmath562th step by induction ( by induction each prime number greater than @xmath384
will be assigned to a prime knot ) .
we call these obtained sequences of composite knots as the preordering sequences of composite knots for constructing the @xmath0th step .
also we call the sequences truncated from these preordering sequences as preordering subsequences of composite knots for constructing the @xmath0th step .
we first have the follwing lemma on preordering sequence .
let @xmath382 be a knot in a preordering sequence of the @xmath0th step .
then there exists a room for this @xmath382 in the @xmath0th step in the sense that this @xmath382 corresponds to a number in the @xmath0th step or in the @xmath562th step .
* proof*. let @xmath382 be of the form @xmath570 where @xmath471 is a knot in the previous @xmath562th step .
by induction we have that @xmath471 is assigned with a number @xmath54 which is the position of @xmath471 in the previous @xmath562th step . then since @xmath571 corresponds to the number @xmath384 we have that @xmath382 corresponds to the number @xmath572 in the @xmath0th step ( we remark that @xmath382 may not be assigned with the number @xmath572 )
thus there exists a room for this @xmath382 in the @xmath0th step
. then let @xmath382 be of the form @xmath573 where @xmath472 is a knot in the previous @xmath563th step .
by induction we have that @xmath472 is assigned with a number @xmath574 which is the position of @xmath472 in the previous @xmath563th step . since @xmath575 is by induction assigned with the prime number @xmath385 we have @xmath576 .
also we have @xmath577 .
thus there exists a room for this @xmath382 in the @xmath562th step or the @xmath0th step .
continuing in this way we have that this lemma holds .
* remark*. by using this lemma we shall construct each @xmath0th step of the classification table by first filling the @xmath0th step with the preordering subsequences of the @xmath0th step .
@xmath72 * remark*. when the number corresponding to the knot @xmath382 in the above proof is not in the @xmath0th step we have that the knot @xmath382 in the preordering sequences of the @xmath0th step has the function of pushing a knot @xmath578 out of the @xmath0th step where this knot @xmath578 is related to a number in the @xmath0th step in order for the knot @xmath382 to be filled into the @xmath0th step .
as an example in the above table the knot @xmath579 ( related to the number @xmath580 ) in a preordering sequence of the @xmath566th step pushes the knot @xmath581 related to the number @xmath582 in the @xmath566th step out of the @xmath566th step .
this relation of pushing out is by the chain @xmath583 . as another example in the above table
the knot @xmath584 ( correspoded to the number @xmath585 ) in a preordering sequence of the @xmath566th step pushes the knot @xmath586 related to the number @xmath587 in the @xmath566th step out of the @xmath566th step .
this relation of pushing out is by the chain @xmath588 .
@xmath72 for @xmath589 the preordering subsequences for the @xmath0th step can cover the whole @xmath0th step .
* proof*. for @xmath590 we have one preordering sequence with number of knots @xmath591 which is obtained by the prime knot @xmath534 taking @xmath524 operation with the step @xmath592 .
in addition we have the knot @xmath593 which is assigned at the position of @xmath594 by the induction procedure .
then since the total rooms of this step @xmath590 is @xmath595 we have that these two knots cover this step @xmath590 . for @xmath596
we have one preordering sequence with number of knots @xmath597 which is obtained by the prime knot @xmath534 taking @xmath524 operation with the step @xmath598 .
this sequence cover half of this step @xmath596 which is with @xmath599 rooms
. then we have one more preordering sequence which is obtained by the knot @xmath479 taking @xmath524 operation with step @xmath569 giving the number @xmath600 of knots .
this covers half of the remaining rooms of the step @xmath596 which is with @xmath601 rooms .
then in addition we have the knot @xmath593 which is assigned at the position of @xmath594 by the induction procedure .
the total of these four knots thus cover the step @xmath596 . for the @xmath0th step we have one preordering sequence with the number of knots @xmath602 which is obtained by the prime knot @xmath534 taking @xmath524 operation with the @xmath537th step .
this sequence cover half of this @xmath0th step which is with @xmath603 rooms .
then we have a preordering sequence which is obtained by the knot @xmath479 taking @xmath524 operation with the @xmath563th step giving the number @xmath604 of knots .
this covers half of the remaining rooms of the @xmath0th step which is with the remaining @xmath605 rooms .
then we have one preordering sequence obtained by picking a prime knot ( e.g.@xmath564 ) which by induction is assigned with a prime number ( e.g. the number 5 ) taking @xmath524 operation with the @xmath568th step .
continue in this way until the knot @xmath536 is by induction assigned at the position of @xmath535 .
the total number of these knots is @xmath603 and thus cover this @xmath0th step .
this proves the lemma . @xmath72
* remark*. since there will have more than one prime number in the @xmath162th steps ( @xmath606 ) in the covering of the @xmath0th step there will have knots from the preordering sequences in repeat and in overlapping . these knots in repeat and in overlapping may be deleted when the ordering of the subsequences of the preordering sequences has been determinated for the covering of the @xmath0th step .
also in the preordering sequences some knots which are in repeat and are not used for the covering of the @xmath0th step will be omitted when the ordering of the subsequences of the preordering sequences has been determinated for the covering of the @xmath0th step .
let us then introduce another definition for constructing the classification table of knots .
* definition ( jumping over of the first kind)*. at an induction @xmath0th step consider a knot @xmath578 and the knot @xmath607 which is a @xmath524 product of @xmath0 knots @xmath608 .
@xmath578 is said to jump over @xmath382 , denoted by @xmath609 , if exist @xmath472 and @xmath544 such that @xmath610 and for any @xmath611 , @xmath471 such that @xmath612 where @xmath611 , @xmath471 , @xmath472 and @xmath544 are not equal to @xmath534 we have @xmath613 or vice versa @xmath614 where @xmath615 , @xmath616 are the numbers assigned to @xmath611 and @xmath471 respectively ( @xmath617 ) and @xmath618 where @xmath619 , @xmath620 are prime knots which have been assigned with prime integers @xmath621 , @xmath622 respectively ; and the following inequality holds : @xmath623 let us call this definition as the property of jumping over of the first kind .
@xmath72 we remark that the definition of jumping over of the first kind is a generalization of the above ordering of @xmath624 and @xmath625 in the above table in the step @xmath626 of @xmath627 .
let us consider some examples of this definition .
consider the knots @xmath628 and @xmath629 .
for any @xmath611 , @xmath471 which are not equal to @xmath534 such that @xmath630 we have @xmath631 and @xmath632 ( or vice versa ) where @xmath385 , @xmath566 are the numbers of @xmath479 and @xmath564 respectively .
thus we have that @xmath633 .
as another example we have that @xmath634 , @xmath635 , and @xmath636 . * a remark on notation*. at the @xmath0th step let a composite knot of the form @xmath637 where
each @xmath638 is a prime knot such that @xmath638 is assigned with a prime number @xmath621 in the previous @xmath537 steps .
then in general @xmath637 is not assigned with the number @xmath639 .
however with a little confusion and for notation convenience we shall sometimes use the notation @xmath640 to denote the knot @xmath637 and we say that this knot is related to the number @xmath640 ( as similar to the knot @xmath571 which is related to the number @xmath384 but is assigned with the number @xmath383 ) and we keep in mind that the knot @xmath637 may not be assigned with the number @xmath640 . with this notation then we may say that the composite number @xmath580 jumps over the number @xmath627 which means that the composite knot @xmath641 jumps over the knot @xmath642 .
@xmath72 * definition ( jumping over of the general kind)*. at the @xmath0th step let a composite knot @xmath578 be related with a number @xmath643 where the number @xmath643 is in the @xmath0th step .
then we say that the knot @xmath578 ( or the number @xmath643 ) is of jumping over of the general kind ( with respect to the knot @xmath382 in the definition of the jumping over of the first kind and we also write @xmath644 ) if @xmath382 satisfies one of the following conditions : \1 ) @xmath578 ( or the number related to @xmath578 ) is of jumping over of the first kind ; or \2 ) there exists a @xmath621 ( for simplicity let it be @xmath645 ) and a prime number @xmath646 such that @xmath645 and @xmath646 are in the same step @xmath162 for some @xmath162 and @xmath646 is the largest prime number in this step such that the numbers @xmath643 and @xmath647 are also in the same step and that the knot @xmath648 related with @xmath647 is of jumping over of the first kind .
@xmath72 * remark*. the condition 2 ) is a natural generalization of 1 ) that if @xmath578 and the knot @xmath648 are as in 2 ) then they are both in the preordering sequences of an induction @xmath0th step or both not . then since @xmath648 is of jumping over into an @xmath649th induction step and
thus is not in the preordering sequences of the induction @xmath0th step we have that @xmath578 is also of jumping over into this @xmath649th induction step ( even if @xmath578 is not of jumping over of the first kind ) .
this means that @xmath578 is of jumping over of the general kind .
@xmath72 * example of jumping over of the general kind*. at an induction step let @xmath578 be represented by @xmath650 ( where we let @xmath651 ) and let @xmath648 be represented by @xmath652 ( where we let @xmath653 )
. then @xmath648 is of jumping over of the first kind .
thus we have that @xmath578 is of jumping over ( of the general kind ) .
@xmath72 we shall show that if @xmath654 then we can set @xmath655 .
thus we have , in the above first example , @xmath656 while @xmath657 . from this property we shall have rooms for the introduction of the @xmath529 operation such that composite numbers are assigned to composite knots and prime numbers are assigned to prime knots .
we have the following theorem . if @xmath658 then it is consistent with the preordering property that @xmath659 for setting up the table . for proving this theorem
let us first prove the following lemma .
the preordering sequences for the construction of the @xmath0th step do not have knots of jumping over of the general kind .
* proof of the lemma*. it is clear that the preordering sequence obtained by the @xmath534 taking a @xmath524 operation with the previous @xmath562th step has no knots with the jumping over of the first kind property since @xmath534 is corresponded with the number @xmath384 and the previous @xmath562th step has no knots with the jumping over of the first kind property for this @xmath562th step .
then preordering sequence obtained by the @xmath479 taking a @xmath524 operation with the previous @xmath563th step has no knots with the jump over of the first kind property since @xmath479 is assigned with the number @xmath385 and @xmath660 and the previous @xmath563th step has no knots with the jumping over of the first kind property for this @xmath563th step . continuing in this way we have that all the knots in these preordering sequences do not satisfy the property of jumping over of the first kind .
then let us show that these preordering sequences have no knots with the property of jumping over of the general kind .
suppose this is not true .
then there exists a knot with the property of jumping over of the general kind and let this knot be represented by a number of the form @xmath643 as in the definition of jumping over of the general kind such that there exists a prime number @xmath646 and that @xmath645 and @xmath646 are in the same step @xmath162 for some @xmath162 and @xmath646 is the largest prime number in this step such that the numbers @xmath643 and @xmath647 are also in the same step and the knot @xmath661 represented by @xmath647 is of jumping over of the first kind . then since @xmath645 and @xmath646 are in the same step @xmath162 we have that the two knots represented by @xmath643 and @xmath647 are elements of two preordering sequences for the construction of the same @xmath0th step .
now since we have shown that the preordering sequences for the construction of the @xmath0th step do not have knots of jumping over of the first kind we have that this is a contradiction .
this proves the lemma . @xmath72
* proof of the theorem*. by the above lemma if @xmath662 then @xmath663 is not in the preordering sequences for the @xmath0th step and thus is pushed out from the @xmath0th step by the preordering sequences for the @xmath0th step and thus we have @xmath664 , as was to be proved .
* remark*. we remark that there may exist knots ( or numbers related to the knots ) which are not in the preordering sequences and are not of jumping over .
an example of such special knot is the knot @xmath665 related with @xmath666 ( but is not assigned with this number ) . @xmath72
* definition*. when there exists a knot which is not in the preordering sequences of the @xmath0th step and is not of jumping over we put this knot back into the @xmath0th step to join the preordering sequences for the filling and covering of the @xmath0th step .
let us call the preordering sequences together with the knots which are not in the preordering sequences of the @xmath0th step and are not of jumping over as the generalized preordering sequences ( for the filling and covering of the @xmath0th step ) .
@xmath72 * remark*. by using the generalized preordering sequences for the covering of the @xmath0th step we have that the knots ( or the number related to the knots ) in the @xmath0th step pushed out of the @xmath0th step by the generalized preordering sequences are just the knots of jumping over ( of the general kind ) .
@xmath72 then we also have the following theorem . at each @xmath0th step ( @xmath667 ) in the covering of the @xmath0th step ( @xmath667 ) with the generalized preordering sequences there are rooms for introducing new knots with the @xmath529 operations .
[ times ] * proof*. we want to show that at each @xmath0th step ( @xmath667 ) there are rooms for introducing new knots with the @xmath529 operations . at @xmath626
we have shown that there is the room at the position @xmath668 for introducing the knot @xmath669 with the @xmath529 operation .
let us suppose that this property holds at an induction step @xmath537 .
let us then consider the induction step @xmath0 . for each @xmath0 because of the relation between @xmath383 and @xmath384 for @xmath670 as a part of the induction step @xmath0 the number @xmath535 is assigned to the knot @xmath536 which is a @xmath524 product of @xmath0 @xmath534 .
then we want to show that for this induction step @xmath0 by using the @xmath671 property we have rooms for introducing the @xmath529 operation .
let @xmath578 be a knot such that @xmath672 and @xmath610 is as in the definition of @xmath671 of jumping over of the first kind such that @xmath673 ( e.g. for @xmath674 we have @xmath675 and @xmath676 ) .
then let us consider @xmath677 .
clearly we have @xmath678 .
thus for each @xmath578 we have a @xmath679 such that @xmath678 .
clearly all these @xmath680 are different .
then from @xmath578 let us construct more @xmath679 , as follows .
let @xmath578 be a knot of jumping over of the first kind .
let @xmath681 and @xmath682 be as in the definition of jumping over of the first kind .
then as in the definition of jumping over of the first kind ( w.l.o.g ) we let @xmath683 then we have @xmath684 also it is trivial that we have @xmath685 and @xmath686 .
this shows that @xmath687 where @xmath688 denotes the knot with the number @xmath689 as in the definition of jumping over of the first kind ( we remark that this @xmath679 corresponds to the knot @xmath690 in the above induction step where @xmath691 is with the number @xmath692 ) .
it is clear that all these more @xmath679 are different from the above @xmath679 constructed by the above method of taking a @xmath524 operation with @xmath534 .
thus there are more @xmath680 than @xmath578 .
thus at this @xmath0th step there are rooms for introducing new knots with the @xmath529 operations .
this proves the theorem . @xmath72
* remark*. in the proof of the above theorem we have a way to construct the knots @xmath680 by replacing a number @xmath54 with the number @xmath693 .
there is another way of constructing the knots @xmath680 by replacing a number @xmath574 with the number @xmath694 . for this way
we need to check that the number related to @xmath680 is in the @xmath562th step for @xmath680 of jumping over into the @xmath0th step . as an example
let us consider the knot @xmath695 of jumping over into the @xmath527th step with the following data : @xmath696 from this data we have : @xmath697 this data gives a knot @xmath680 with the related number @xmath698 . on the other hand from the data ( [ jumpingover9 ] ) we have
: @xmath699 since @xmath700 we have that the knot @xmath701 related with the number @xmath702 is of jumping over into the @xmath567th step ( we shall show that @xmath565 is assigned with the number @xmath567 ) . @xmath72 * remark*. the above theorem shows that at each @xmath0th step there are rooms for introducing new knots with the @xmath529 operations and thus we may establish a one - to - one correspondence of knots and numbers such that prime knots are bijectively assigned with prime numbers .
further to this theorem we have the following main theorem : a classification table of knots can be formed ( as partly described by the above table up to @xmath535 with @xmath703 ) by induction on the number @xmath535 such that knots are one - to - one assigned with an integer and prime knots are bijectively assigned with prime numbers such that the prime number @xmath384 corresponds to the trefoil knot .
this assignment is onto the set of positive integers except @xmath384 where the trefoil knot is assigned with 1 and is related to @xmath384 and at each @xmath0th induction step of the number @xmath535 there are rooms for introducing new knots with the @xmath529 operations only .
further this assignment of knots to numbers for the @xmath0th induction step of the number @xmath535 effectively includes the determination of the distribution of prime numbers in the @xmath0th induction step and is by induction determined by this assignment for the previous @xmath537 induction steps such that the assignment for the previous @xmath537 induction steps is inherited in this assignment for the @xmath0th induction step as the preordering sequences in the determination of this assignment for the @xmath0th induction step .
[ maintheorem ] * remark*. let us also call this assignment of knots to numbers as the structure of numbers obtained by assigning numbers to knots .
this structure of numbers is the original number system together with the one - to - one assignment of numbers to knots .
* proof*. by the above lemmas and theorems we have that the generalized preordering sequences have the function of pushing out those composite knots of jumping over from the @xmath0th step .
it follows that for step @xmath667 there must exist chains of transitions whose initial states are composite knots in repeat ( to be replaced by the new composite knots with @xmath529 operations only ) ; or the knots of jumping over into this @xmath0th step from the previous @xmath562th step ; or the knots in the preordering sequences with the @xmath529 operations ; such that the composite knots of jumping over are pushed out from the @xmath0th step by these chains .
these chains are obtained by ordering the subsequences of preordering sequences such that the preordering property holds in the @xmath0th step .
further the intermediate states of the chains must be positions of composite numbers .
this is because that if a chain is transited to an intermediate state which is a position of prime number then there are no composite knots related by this prime number and thus this chain can not be transited to the next state and is stayed at the intermediate state forever and thus the chain can not push out the composite knot of jumping over .
then when a composite knot is at the position of an intermediate state ( which is a position of composite number as has just been proved ) then this knot is definitely assigned with this composite number .
then when a composite knot which is in repeat is at the position of an intermediate state then this knot is also definitely assigned with this composite number .
it follows that when the chains are completed we have that the ordering of the subsequences of preordering sequences is determined .
then the remaining knots ( which are not at the transition states of the chains ) which are not in repeat are definitely assigned with the number of the positions of these knots in the @xmath0th step .
for these knots the numbers of positions assigned to them are just the number related to them respectively
. then the remaining knots ( which are not at the transition states of the chains ) which are in repeat must be replaced by new prime knots because of the repeat and that no other knots related with numbers in this @xmath0th step in the generalized preordering sequences can be used to replace the remaining knots .
this means that the numbers of the positions of these remaining knots in repeat are prime numbers in this @xmath0th step .
this is because that if the numbers of the positions assigned to the new prime knot is a composite number then the composite knot related with this composite number is either in a transition state or is not in transition .
if the composite knot is not in transition then the composite number related to this composite knot is just the number assigning to this composite knot and since this number is also assigned to the new prime knot that this is a contradiction .
then if this composite knot is in transition state then this means that the remaining knot is also in transition state and this is a contradiction since by definition the remaining knot is not at the transition states of the chains .
thus prime numbers in the @xmath0th step are assigned and are only assigned to prime knots which replace the remaining knots in repeat in the @xmath0th step .
thus from the preordering sequences we have determined the positions ( i.e. the distribution ) of prime numbers in the @xmath0th step .
now since the preordering sequences are constructed by the previous steps we have shown that the basic structure ( in the sense of above proof ) of this assignment of knots with numbers for the @xmath0th step ( including the determination of the distribution of prime numbers in the @xmath0th step ) is determined by this assignment of knots with numbers for the previous @xmath537 steps . in other words we have that the basic structure of the @xmath0th induction step is determined by the structure of the previous @xmath537 steps . to complete the proof of this theorem let us show that at each @xmath0th induction step ( @xmath667 ) there are rooms for introducing new composite knots with the @xmath529 operations only and we can determine the ordering of these composite knots with the @xmath529 operations only in each @xmath0th induction step .
in the above proof we have shown that the basic structure of the @xmath0th induction step is determined by the structure of the previous steps such that the positions of the composite knots with the @xmath529 operations only in the @xmath0th induction step are correctedly determined by the structures of the previous steps .
these positions are fitted for the corrected composite knots with the @xmath529 operations only constructed ( by the @xmath529 operations ) by knots in the previous steps .
thus for this @xmath0th induction step the introducing and the ordering of composite knots with the @xmath529 operations only is also determined by the structures of the previous @xmath537 steps .
further since the structures of the previous steps are inherited in the structure of the @xmath0th induction step as the preordering sequences in the determination of the structure of the @xmath0th induction step we have that all the properties of the structures of the previous steps are inherited in the structure of the @xmath0th induction step in the determination of the structure of the @xmath0th induction step .
thus the new composite knots with the @xmath529 operations only in the @xmath0th induction step inherit the ordering properties ( such as the preordering property ) of composite knots with the @xmath529 operations only in the previous steps .
( these ordering properties of the composite knots with the @xmath529 operations only can be used to find out the corrected composite knots with the @xmath529 operations only to be assigned at the corrected positions in the @xmath0th step ) . with this fact let us then show that at each @xmath0th induction step ( @xmath667 ) there are rooms for introducing new composite knots with the @xmath529 operations only . as in the proof of the theorem [ times ]
we first construct more @xmath680 by the method following ( [ forjumpingover ] ) .
let us start at the step @xmath626 .
for this step we have the knot @xmath704 jumps over into the step @xmath703 . for this @xmath578
we have the following data as in ( [ forjumpingover ] ) : @xmath705 from ( [ jumpingover3 ] ) we construct a @xmath680 for the step @xmath703 by the following data : @xmath706 this data gives one more @xmath707 .
then from ( [ jumpingover3 ] ) we construct one more @xmath680 for the step @xmath703 by the following data : @xmath708 this data gives one more @xmath709 .
thus in this step @xmath703 there are two rooms for the two knots @xmath710 and @xmath711 coming from the preordering sequences and there exists exactly one room for introducing a new composite knot with the @xmath529 operations only ( recall that we also have a @xmath712 ) . from the ordering of knots in the previous steps we determine that @xmath713 is the composite knot with the @xmath529 operations only for this step .
thus at the @xmath528th and @xmath566th steps we can and only can introduce exactly one composite knot with the @xmath529 operations only and they are the knots @xmath714 and @xmath715 respectively .
this shows that at the @xmath528th and the @xmath566th steps we can determine the number of prime knots with the minimal number of crossings @xmath716 and @xmath717 respectively ( these two prime knots are denoted by @xmath534 and @xmath479 respectively and we do not distinguish knots with their mirror images for this determination of the ordering of knots with the @xmath529 operations only .
this also shows that there are rooms for introducing new composite knots with the @xmath529 operations only in the @xmath528th and @xmath566th steps ) .
then since this property is inherited in the @xmath527th step we can thus determine that the @xmath527th step is a step for introducing new composite knots with the @xmath529 operations only of the form @xmath718 where @xmath719 denotes a prime knot with the minimal number of crossings @xmath720 ( and thus there are rooms for introducing new composite knots with the @xmath529 operations only in this @xmath527th step ) .
also since the properties in the @xmath528th and @xmath566th steps are inherited in the @xmath527th step we can determine the number of prime knots with the minimal number of crossings @xmath720 by the knots of the form @xmath718 as this is a property of knots with the @xmath529 operations only in the @xmath528th and @xmath566th steps ( in the classification table in the next section we show that there are exactly two composite knots of the form @xmath721 and @xmath722 in the @xmath527th step whose ordering are determined by the preordering property of knots and the structure of the @xmath527th step .
this thus shows that there are exactly two prime knots with the minimal number of crossings @xmath720 and they are denoted by @xmath564 and @xmath565 respectively ) .
then since the properties of the @xmath528th , @xmath566th and @xmath527th steps are inherited in the @xmath567th step we can determine that the @xmath567th step is a step for introducing new composite knots with the @xmath529 operations only of the form @xmath723 where @xmath724 denotes a prime knot with the minimal number of crossings @xmath725 ( and thus there are rooms for introducing new composite knots with the @xmath529 operations only in this @xmath567th step ) .
also since the properties in the @xmath528th , @xmath566th and @xmath527th steps are inherited in the @xmath567th step we can determine the number of prime knots with the minimal number of crossings @xmath725 by the knots of the form @xmath723 as this is a property of knots with the @xmath529 operations only in the @xmath528th , @xmath566th and @xmath527th steps ( in the classification table in the next section we show that there are exactly three composite knots of the form @xmath726 , @xmath727 and @xmath728 in the @xmath567th step whose ordering are determined by the preordering property of knots and the structure of the @xmath567th step .
this thus shows that there are exactly three prime knots with the minimal number of crossings @xmath725 and they are denoted by @xmath729 , @xmath730 and @xmath731 respectively ) . continuing in this way
we thus show that at each @xmath0th induction step @xmath732 we can determine the number of prime knots with the minimal number of crossings @xmath733 and there are rooms for introducing new composite knots with the @xmath529 operations only .
this proves the theorem . @xmath72
* example*. let us consider the above table up to @xmath538 ( with @xmath0 up to @xmath566 ) as an example . for the induction step at @xmath590 ( or at @xmath734 )
we have one preordering sequence obtained by letting @xmath534 to take a @xmath524 operation with the step @xmath569 ( for the step @xmath569 the number @xmath595 is related to the trefoil knot @xmath534 ) : @xmath735 .
then we fill the step @xmath590 with this preordering sequence and we have the following ordering of knots for this step @xmath590 : @xmath736 where the first @xmath737 placed at the position @xmath385 is the preordering sequence while the second @xmath737 placed at the position @xmath734 is required by the induction procedure . for this step
there is no numbers of jumping over .
then we have that the first @xmath737 is a repeat of the second @xmath737 .
thus this repeat one must be replaced by a new prime knot .
let us choose the prime knot @xmath479 to be this new prime knot since @xmath479 is the smallest of prime knots other than the trefoil knot .
then this new prime knot must be at the position of a prime number , as we have proved in the above theorem .
thus we have determined that @xmath385 is a prime number in this step @xmath590 by using the structure of numbers of step @xmath569 which is only with the prime number @xmath384 .
then for the induction step at @xmath596 ( or at @xmath738 ) we have two preordering sequence obtained by letting @xmath479 to take a @xmath524 operation with the step @xmath569 and by letting @xmath534 to take a @xmath524 operation with the step @xmath590 : @xmath739 where the first knot is the preordering sequence obtained by letting @xmath479 to take a @xmath524 operation with the step @xmath569 and the second and third knots is the preordering sequence obtained by letting @xmath534 to take a @xmath524 operation with the step @xmath590 . for this step there is no numbers of jumping over and thus there are no chains of transition .
thus the ordering of the above three knots in this step follow the usual ordering of numbers . thus the number assigned to the knot @xmath740 must be assigned with a number less than that of @xmath741 by the ordering of @xmath742 and @xmath741 in the second preordering sequence . by this ordering of the two preordering sequences we have that the step @xmath596 is of the following form : @xmath743 where the fourth knot @xmath741 is put at the position of @xmath738 and is assigned with the number @xmath738 as required by the induction procedure .
thus the third knot @xmath744 is a repeated one and thus must be replaced by a prime knot and the position of this prime knot is determined to be a prime number .
thus we have determined that the number @xmath567 is a prime number . then since there are no chains of transition we have that the composite knot @xmath742 must be assigned with the number related to this knot and this number is @xmath745 .
thus the composite knot @xmath742 is at the position of @xmath527 and that the first knot @xmath746 is a repeat of the second knot and thus must be replaced by a prime knot .
then since this prime knot is at the position of @xmath566 we have that @xmath566 is determined to be a prime number .
now the two prime knots at @xmath566 and @xmath567 must be the prime knots @xmath564 and @xmath565 respectively since these two knots are the smallest prime knots other than @xmath534 and @xmath479 ( we may just put in two prime knots and then later determine what these two knots will be . if we put in other prime knots then this will not change the distribution of prime numbers determined by the structure of numbers of the previous steps and it is only that the prime knots are assigned with incorrect prime numbers .
further as shown in the above proof by using knots of the form @xmath747 we can determine that there are exactly two prime knots with minimal number of crossings @xmath720 and they are denoted by @xmath564 and @xmath565 respectively .
from this we can then determine that these two prime knots are @xmath564 and @xmath565 ) .
thus we have the following ordering for @xmath596 : @xmath748 where @xmath564 is assigned with the prime number @xmath566 and @xmath565 is assigned with the prime number @xmath567 .
this gives the induction step @xmath596 .
for this step there is no knot with @xmath529 operation since there is no knots of jumping over .
let us then consider the step @xmath626 ( or @xmath627 ) . for this step
we have the following three preordering sequences obtained from the steps @xmath749 : @xmath750 where the third sequence is obtained by taking @xmath524 operation of the knot @xmath534 with step @xmath596 while the third sequence is obtained by taking @xmath524 operation of the knot @xmath479 with the step @xmath590 and the first sequence is obtained by taking @xmath524 operation of the knot @xmath564 with step @xmath569 .
then as required by the induction procedure the knot @xmath751 is assigned at the position of @xmath627 .
the total number of knots in ( [ order2 ] ) plus this knot is exactly @xmath738 which is the total number of this step @xmath626 . * remark*. we have one more preordering sequence obtained by taking @xmath524 operation of the knot @xmath565 with step @xmath569 .
this preordering sequence gives the knot @xmath752 .
however since the knots in ( [ order2 ] ) and the knot @xmath751 assigned at the position of @xmath627 are enough for covering this step @xmath626 and that the knot @xmath752 of this preordering sequence is a repeat of the knot @xmath752 in ( [ order2 ] ) that this preordering sequence obtained by taking @xmath524 operation of the knot @xmath565 with step @xmath569 can be omitted .
@xmath72 then to find the chains of transition for this step let us order the three preordering sequences with the following ordering where we rewrite the preordering sequences in column form and the knot @xmath751 assigned at the position of @xmath627 is put to follow the three sequences : @xmath753 we notice that this column exactly fills the step @xmath626 .
for this step we have that the number @xmath580 ( or the knot @xmath754 related with @xmath580 ) is of jumping over . from ( [ step4 ] )
we have the following chain of transition for pushing out @xmath754 at @xmath580 by a knot with the @xmath529 operation replacing the repeated knot @xmath755 at the position of @xmath756 : @xmath757 where we choose the knot @xmath669 as the knot with the @xmath529 operation since @xmath669 is the smallest one of such knots . for this chain
the intermediate states are at positions of composite numbers @xmath758 , @xmath759 and @xmath580 .
thus the knots in this chain at the positions of these composite numbers are assigned with these composite numbers respectively .
then once this chain of pushing out @xmath754 at @xmath580 is set up we have that the other knots in repeat must by replaced by prime knots and that their positions must be prime numbers .
these positions are at @xmath760 and @xmath761 and thus @xmath760 and @xmath761 are determined to be prime numbers ( the knot @xmath762 at the end of this step must be assigned with @xmath763 by the induction procedure and thus the knot at @xmath761 is a repeat ) . then the new prime knots @xmath729 and @xmath730 are suitable knots corresponding to the prime numbers @xmath760 and @xmath761 respectively since they are the smallest prime knots other than @xmath534 .
@xmath479 , @xmath564 and @xmath565 ( as the above induction step we may just put in two prime knots and then later determine what these two prime knots will be . as shown in the above proof by using knots of the form @xmath764 we can determine that there are exactly three prime knots with minimal number of crossings @xmath725 and they are denoted by @xmath729 , @xmath730 and @xmath731 respectively . from this
we can then determine that these two prime knots are @xmath729 and @xmath730 ) .
this completes the step @xmath626 .
thus the structure of numbers of this step ( including distribution of prime numbers in this step ) is determined by the structure of numbers of the previous induction steps .
let us then consider the step @xmath703 . for this step
we have the following four preordering sequences from the previous steps @xmath765 : @xmath766 and @xmath767 and @xmath768 and @xmath769 the total number of knots ( including repeat ) in the above sequences plus the knot @xmath770 to be assigned at the position of @xmath538 exactly cover this @xmath703 step . * remark*. as similar to the step @xmath626 two preordering sequences @xmath771 and @xmath772 are omitted since these sequences are with knots which are repeats of the knots in the above preordering sequences .
@xmath72 then to find the chains of transition for this step let us order these four preordering sequences with the following ordering where the knot @xmath770 assigned at the position of @xmath538 is put to follow the four sequences : @xmath773 for this step we have three composite knots @xmath774 , @xmath775 and @xmath776 ( related with @xmath587,@xmath582 and @xmath777 respectively ) of jumping over and there are two new knots @xmath754 and @xmath778 coming from the previous step . thus there is a room for the introduction of new knot obtained only by the @xmath529 operation .
then this new knot must be the composite knot @xmath715 since besides the composite knot @xmath714 it is the smallest of composite knots of this kind . from ( [ cc5 ] )
there is a chain of transition given by @xmath779 and the composite knot @xmath776 related with @xmath780 is pushed out into the next step by the composite knot @xmath781 at the starting position @xmath782 .
then this repeated knot must be replaced by a new composite knot obtained by the @xmath529 operation only and this new composite knot must be the knot @xmath715 .
then the composite knots at the intermediate states are assigned with the numbers of these states respectively .
in addition to the above chain there are two more chains : @xmath783 and @xmath784 .
the chain @xmath783 starts from @xmath778 at @xmath785 and the composite knot @xmath774 at @xmath786 is pushed out by the composite knot @xmath787 .
then the chain @xmath784 starts from @xmath754 at @xmath788 and the composite knot @xmath775 at @xmath789 is pushed out by the composite knot @xmath790 .
then the knots @xmath787 and @xmath790 at the intemediate states of these two chains are assigned with the numbers @xmath791 and @xmath792 respectively .
now the remaining repeated composite knots at the positions @xmath793 must be replaced by new prime knots and thus @xmath793 are determined to be prime numbers and they are determined by the prime numbers in the previous induction steps . then we may follow the usual table of knots to determine that the new prime knots for the prime numbers @xmath793 are @xmath731 , @xmath794 , @xmath795 , @xmath796 and @xmath797 respectively ( as the above induction steps we may just put in five prime knots and then later determine what these five prime knots will be . as shown in the above proof by using knots of the form @xmath798 we can determine the number of prime knots with minimal number of crossings @xmath799 . from this we can then determine these five prime knots ) . in summary
we have the following form of the step @xmath703 : @xmath800 this completes the induction step at @xmath703 .
we have that the structure of numbers of this step ( including distribution of prime numbers in this step ) is determined by the structure of numbers of the previous induction steps . @xmath72
following the above classification table up to @xmath538 let us in this section give the table up to @xmath801 .
again we shall see from the table that the preordering property is clear . at the @xmath567th step there is
a special composite knot @xmath802 which is not of jumping over and is not in the preordering sequences ( on the other hand the knot @xmath803 is of jumping over ) .
we remark again that it is interesting that ( by the ordering of composite knots with the @xmath529 operation only ) at the @xmath527th step we require exactly two prime knots with minimal number of crossings @xmath720 to form the two composite knots obtained by the @xmath529 operation only . from this
we can determine the number of prime knots with minimal number of crossings @xmath720 without using the actual contruction of these prime knots .
we then denote these two prime knots by @xmath564 and @xmath565 respectively and the two composite knots obtained by the @xmath529 operation only by @xmath804 and @xmath805 respectively . similarly at the @xmath567th step
we can determine that there are exactly three prime knots with minimal number of crossings @xmath725 and we denote these three prime knots by @xmath729 and @xmath730 and @xmath731 respectively .
these three prime knots give the composite knots @xmath806 , @xmath807 and @xmath808 respectively .
we can then expect that at the next @xmath809th step we may determine that the number of prime knots with minimal number of crossings @xmath799 is @xmath567 and then at the next @xmath668th step the number of prime knots with minimal number of crossings @xmath810 is @xmath420 , and so on ; as we know from the well known table of prime knots @xcite . here
the point is that we can determine the number of prime knots with the same minimal number of crossings without using the actual construction of these prime knots ( and by using only the classification table of knots ) . @xmath811
@xmath812 @xmath813
let us extend the above quantum knots and knot invariant to quantum links and link invariant .
let us first consider some examples to see how the quantum link and link invariant is defined .
let us consider the link in fig.6a .
we may let the two knots of this link be with @xmath70 and @xmath244 as the initial and final end point respectively .
we let the ordering of these two knots be such that when the @xmath38 parameter goes one loop on one knot then the @xmath38 parameter for another knot also goes one loop .
the trace invariant ( [ m14 ] ) for this link is given by : @xmath814 we let the ordering of the wilson lines in ( [ l1a ] ) be such that @xmath2 and @xmath815 start first . then next @xmath816 and @xmath3 follows .
form this ordering we have that ( [ l1a ] ) is equal to : @xmath817 where we have used ( [ m7a ] ) and ( [ m8a ] ) .
since by definition ( [ m14 ] ) we have that @xmath818 is the knot invariant for two unlinking trivial knots , equation ( [ l1 ] ) shows that the link in fig.6a is topologically equivalent to two unlinking trivial knots .
similarly we can show that the link in fig.6b is topologically equivalent to two unlinking trivial knots .
fig.6a fig.6b fig.7a fig.7b let us then consider the hopf link in fig.7a .
the trace invariant ( [ m14 ] ) for this link is given by : @xmath819 the ordering of the wilson lines in ( [ l3 ] ) is such that @xmath2 starts first and @xmath3 follows it .
then next we let @xmath816 starts first and @xmath815 follows it .
let us call this ordering as the simultaneous ordering which will be used to define the braiding formulas for a crossing between two knot components of a link .
this ordering has a property that when the @xmath38 parameter has traced one loop in one knot of the link we have that the @xmath38 parameter has also traced one loop on the other knot .
from this ordering we have that ( [ l3 ] ) is equal to : @xmath820 then let us consider the following trace : @xmath821 we let the ordering of the wilson lines in ( [ l5 ] ) be such that @xmath2 starts first and @xmath815 follows it .
then next @xmath816 starts first and @xmath3 follows it . from this ordering
we have that ( [ l5 ] ) is equal to : @xmath822 then since in ( [ l6 ] ) the crossings between the two knot components have been eliminated we have that the two knot components are independent and thus the starting points for the two knot components are independent and thus ( [ l6 ] ) is equal to ( [ l4 ] ) .
on the other hand from the ordering of ( [ l5 ] ) we have that ( [ l5 ] ) is equal to : @xmath823 where we have repeatedly used ( [ m9 ] ) . from ( [ l4 ] ) , ( [ l6 ] ) and ( [ l7 ] ) we have that the knot invariant for the hopf link in fig.7a is given by : @xmath824 we remark that in ( [ l8 ] ) since @xmath7 is a @xmath7-matrix between two knot components of the hopf link we have that @xmath7 acts on @xmath825 or on @xmath826 .
in this case we say that the domain of @xmath7 is @xmath827 . from this property of
@xmath7 we have that the @xmath7 and the monodromies @xmath828 for @xmath829 and @xmath830 in ( [ l8 ] ) are independent .
then let us consider the hopf link in fig.7b .
the correlation for this link is given by @xmath831 by a derivation which is dual to the above derivation for the hopf link in fig.7a we have that ( [ l9 ] ) is equal to @xmath832 where the @xmath7 and the monodromies for @xmath833 and @xmath364 in ( [ l10 ] ) are independent .
we see that the invariants for the above two hopf links are different .
this agrees with the fact that these two links are not topologically equivalent . as more examples let us consider the linking of two trivial knots with linking number @xmath384 as in fig.8a ( the reader may skip the following of this section for the first reading ) .
similar to the above computations we have that this link which is analogous to the hopf link in fig.7a is with an invariant equals to @xmath834 . also for the link in fig.8b which is analogous to the hopf link in fig.7b is with an invariant equals to @xmath835 .
fig.8a fig.8b let us give a computation of the link invariant of the link diagram in fig.8a , as follows .
by definition the link invariant of this diagram is the trace of the generalized wilson loop of this link diagram which is given by : @xmath836 where the ordering is such that @xmath2 stars first and @xmath837 follows it . then next we let @xmath838 stars first and @xmath839 follows it .
continuing in this way we have an ordering such that when the @xmath38 parameter has traced one loop we have that the @xmath840 parameter has also traced one loop . from the ordering and the braiding formulas ( [ m7a ] ) , ( [ m8a ] )
we have that ( [ link ] ) is equal to : @xmath841 on the other hand as similar to the hopf link let us consider the following trace : @xmath842 where we let the ordering be such that @xmath2 starts first and @xmath843 follows it .
continuing in this way the ordering in ( [ link3 ] ) is then determined . from this ordering
we have that ( [ link3 ] ) is equal to : @xmath844 then since the two knot components are independent we have that the starting points for the two knot components are independent and we thus have that ( [ link4 ] ) is equal to ( [ link2 ] ) .
on the other hand from the ordering of ( [ link3 ] ) we have that ( [ link3 ] ) is equal to : @xmath845 where the final step is from the above derivation of the invariant of the hopf link .
this shows that the invariant of the knot diagram ( a ) in fig.8 is equal to @xmath846 where @xmath7 is independent of the monodromies of @xmath847 and @xmath364 .
similarly we can show that the invariant of the knot diagram ( b ) in fig.8 is equal to @xmath848 .
let us generalize the hopf link to the case with linking number @xmath0 .
then by induction on the above results we have that the two generalized hopf links with linking number @xmath0 are respectively with invariants @xmath849
similar to the case of knot for each link @xmath10 let us construct the generalized wilson loop @xmath850 . for the case of link in constructing the generalized wilson loop we need to consider the crossings between two knot components of a link . as shown in the hopf link example for a crossing between two knot components of a link we give it a simultaneous ordering such that the braiding formulas for such crossing are defined .
when the braiding formulas are defined we have then completely represented this crossing by its wilson product .
once a crossing between two knot components of a link @xmath10 is completely represented by its wilson product we can then follow the orientations of each knot component of this link @xmath10 to write out the sequence of wilson products for the sequence of crossings on each knot component of the link . by writing out all these sequences of wilson products of each knot component one by one
until all crossings have been counted we have that the generalized wilson loop @xmath850 of @xmath10 is then formed . in this process of counting the crossings we have that the crossings which have been counted once will not been counted again when they reappear . from these reappearances
we have the property of circling and sub - circling of the link .
let us consider some examples to illustrate the construction of @xmath850 .
as a simple example let us consider again the hopf links in fig.7 .
we let an ordering be such that @xmath2 starts first and @xmath3 follows it simultaneously .
this is by definition a simultaneous ordering of @xmath2 and @xmath3 .
then next we let @xmath816 starts first and @xmath815 follows it simultaneously .
this is by definition a simultaneous ordering of @xmath816 and @xmath815 . for the hopf link if we let @xmath383 denote the crossing of @xmath2 with @xmath3 and let @xmath384 denote the crossing of @xmath816 with @xmath815 .
then we have @xmath851 .
let us also denote the corresponding crossings of the hopf link by @xmath383 and @xmath384 respectively .
then for the hopf link we have the following circling property : @xmath852 then by using exactly the same method for proving the circling property of @xmath398 of a knot @xmath382 we can show that the generalized wilson loop @xmath851 of the hopf link @xmath10 also has the circling property ( [ o3 ] ) . in the following
let us consider more examples of @xmath850 and the circling property of links .
* examples of @xmath850 and the circling property of links*. as an example let us consider the link @xmath10 in fig.9
. fig.9 let @xmath47 denote the crossing at @xmath853 .
then the generalized wilson loop @xmath850 is given by @xmath854 .
then by using exactly the same method for proving the circling property of @xmath398 of a knot @xmath382 we can show that the generalized wilson loop @xmath854 satisfies the following three circlings of @xmath10 : @xmath85534 = [ 21]34 = [ 12]34 = \cdot\cdot\cdot \\ & 1234 = 4123 = 3412 = 2341 = 1234 = \cdot\cdot\cdot \\ & 12[34 ] = 12[43 ] = 12[34 ] = \cdot\cdot\cdot \end{array } \label{o4}\ ] ] where the sequences in the bracket @xmath856 $ ] form a circling . as one more example let us consider the borromeo ring @xmath10 in fig.10 .
fig.10 let @xmath47 denote the crossing at @xmath857 .
then we have many ways to write out @xmath10 which are all equivalent by the circling property of @xmath10 . as an example we have the following three circlings of @xmath10 where each sequence of crossings representing @xmath10 : @xmath8585(6)3 = [ 2164]5(6)3 = [ 4216]5(6)3 = [ 6421]5(6)3=[1642]5(6)3 = \cdot\cdot\cdot \\ & [ 2563]4(5)1 = [ 3256]4(5)1 = [ 6325]4(5)1 = [ 5632]4(5)1
= [ 2563]4(5)1= \cdot\cdot\cdot \\ & [ 3451]6(4)2 = [ 1345]6(4)2 = [ 5134]6(4)2 = [ 4513]6(4)2 = [ 3451]6(4)2 = \cdot\cdot\cdot \end{array } \label{o4a}\ ] ] in the above sequences the notation @xmath390 means that the number @xmath352 is circled to the position of @xmath352 without @xmath859 as indicated in the sequences ( also this notation @xmath390 means that the @xmath352 in @xmath859 reappears and is not counted ) .
also the above three circlings of sequences can be circled to each other .
for example we have the following circling : @xmath8605(6)3=[1(6)42]563 = 1(6)4[2563 ] = 14[2563 ] = 4[2563]1 \\
= & [ 2563]14=[2563]1(3)4(5 ) = [ 2563]4(5)1(3 ) = [ 2563]4(5)1=\cdot\cdot\cdot \end{array } \label{o5}\ ] ] this shows that the first and the second circlings of the above three circlings can be circled to each other .
then each of the above sequences can be as the generalized wilson loop @xmath850 and by using exactly the same method for the case of knot for proving the cirling property we can show that the generalized wilson loop @xmath850 also has the above circling properties ( [ o4 ] ) and ( [ o5 ] ) .
@xmath72 in general by using exactly the same method for the case of knot for proving the cirling property of the generalized wilson loop of a knot we can show that the generalized wilson loop of a general link also has the above circling property of this general link . with this circling property of the generalized wilson loop as similar to the case of knot
we have the following theorem for links : each link @xmath10 can be faithfully represented by its generalized wilson loop @xmath850 in the sense that if two link diagrams have the same generalized wilson loop then these two link diagrams must be equivalent . from this theorem on faithful representation of links we have the following classification theorem for links .
let @xmath850 denote the generalized wilson loop of a link @xmath10 with component knots @xmath861 .
then @xmath850 is a link invariant which classifies links .
we have that @xmath850 can be written in the following form : @xmath862 where @xmath863 are monodromies of the kz equation which come from the linkings of @xmath864 where the domains of @xmath865 are subsets of @xmath866 .
also these monodromies @xmath865 and monodromies for @xmath867 are independent of each other in the sense that the acting domains of these monodromies are different .
then the trace @xmath868 is also a link invariant which classifies links .
* proof . *
the proof of this theorem is similar to the case of knot . as similar to the case of knot
let us first find the following expression for @xmath869 : @xmath870 where @xmath865 are independent monodromies which are also independent of the monodromies for @xmath871 . from this expression and the above theorem on faithful representation of link
we then have that the link invariant @xmath868 classifies links .
let @xmath10 be a trivial link with two unlinking component @xmath471 and @xmath472 .
we may suppose that @xmath471 and @xmath472 have no crossings to each other .
then @xmath850 is in the following form : @xmath872 where we have @xmath873 for some @xmath874 .
we have that the @xmath875 matrices operating on two independent operators @xmath296 and @xmath301 respectively .
let @xmath471 and @xmath472 be linked to form a link @xmath10 .
then from the theorem in the section on solving the kz equations we have that @xmath850 is in a tensor product form . since @xmath471 and @xmath472 are two closed curves we have that this tensor product reduces its degree to a product with a tensor product of the form @xmath876 where @xmath296 and @xmath301 are two independent operators for @xmath471 and @xmath472 respectively .
then since the matrices @xmath877 and @xmath878 act on either @xmath296 or @xmath301 we have that they commute with @xmath876 and thus we can write @xmath850 in the following form : @xmath879 where the monodromies @xmath880 acts either on @xmath296 or @xmath301 . since @xmath881 for some @xmath882 from ( [ cl ] ) we can write @xmath850 in the following form : @xmath883 where @xmath884 .
we have that the monodromies @xmath885 in ( [ cl1 ] ) must be independent of the monodromies @xmath875 since if @xmath886 then it will be absorbed by @xmath871 to form a generalized wilson loop @xmath887 for some knot @xmath888 which is not equivalent to @xmath871 .
this is impossible since @xmath10 is not formed with this knot @xmath888 . on the other hand the monodromies @xmath885 in ( [ cl1 ] )
can be set to be independent of each other ( in the sense that the acting domains of these monodromies are different ) since if two @xmath885 are the same then they can be merged into one @xmath885 .
this form ( [ cl1 ] ) of @xmath850 is just the required form ( [ aa3a ] ) . for a general @xmath10
the proof of this form of @xmath850 is similar .
this proves the theorem . @xmath72
* remark*. let us give more details on the domain of a monodromy , as follows .
for simplicity let us consider the above link @xmath10 with two components @xmath471 and @xmath472 .
we have that a monodromy @xmath885 acts on @xmath296 or @xmath301 ( or acts on @xmath475 or @xmath476 .
thus the domain of @xmath885 is actually a subset of @xmath889 .
let us consider the hopf link as a simple example .
for the hopf link @xmath10 we have that @xmath890 . in this
@xmath850 the monodromy @xmath7 is with domain @xmath889 since @xmath7 is obtained by braiding between the wilson lines of @xmath471 and @xmath472 . on the other hand
the @xmath875 are for the forming of @xmath871 . in this case
we say that @xmath882 are with domains @xmath891 respectively .
in this section we derive quantum invariants of closed 3-manifolds from the above quantum invariants of links .
we have the lickorish - wallace theorem which states that any closed ( oriented and connected ) 3-manifold @xmath9 can be obtained from a dehn surgery on a framed link @xmath10 @xcite@xcite@xcite@xcite .
let us first consider 3-manifolds obtained from surgery on framed knots @xmath892 where @xmath893 and @xmath646 are co - prime integers .
we have the following expression of the generalized wilson loop @xmath894 of @xmath892 : @xmath895 where the @xmath7-matrix @xmath7 denotes the linking matrix which acts on @xmath896 where @xmath897 denotes the partner ( or company ) of @xmath382 for the framed knot @xmath892 . in ( [ q11a ] )
we write @xmath898 and @xmath899 .
the integers @xmath464 and @xmath900 are indexes for the knot @xmath382 and its partner @xmath897 respectively .
on the other hand in ( [ q11a ] ) the @xmath7-matrix @xmath901 acting on @xmath398 and @xmath902 is from the linking of @xmath382 and @xmath897 ( and from the number @xmath646 ) and is for the effect of giving @xmath903 linking number .
then since @xmath897 is as the partner of @xmath382 in the construction of @xmath9 we have that the @xmath7-matrices @xmath865 acting on @xmath904 @xmath905 respectively are such that @xmath906 .
from this we have that @xmath907 and @xmath906 are as the same function on @xmath829 and @xmath830 and thus @xmath908 .
thus from ( [ q11a ] ) we have the following representation of @xmath9 : @xmath909 where we have absorbed @xmath901 into the matrices @xmath865 @xmath905 and for simplicity the resulting indexes are still denoted by @xmath464 and @xmath900 respectively ( for simplicity we still use @xmath398 to denote @xmath910 for the resulting index @xmath464 which may be different from the original @xmath464 for @xmath398
. similarly we still use @xmath902 to denote @xmath911 for the resulting index @xmath900 which may be different from the original @xmath900 for @xmath902 ) .
thus in the case of the linking of @xmath382 and @xmath897 giving the effect of @xmath903 linking number there may have many surgeries on different @xmath382 ( with different original @xmath464 and @xmath900 but giving the same resulting indexes @xmath464 and @xmath900 ) giving the same @xmath9 by this degeneration @xcite .
we notice that all these surgeries are with the same representation ( [ q11 ] ) of @xmath9 .
then from kirby calculus @xcite we have that ( [ q11 ] ) may still be a many - to - one representation of 3-manifold @xmath9 obtained from surgery on @xmath892 .
let us from ( [ q11 ] ) find a one - to - one representation ( or invariant ) of 3-manifold @xmath9 .
to this end let us investigate homeomorphisms ( or symmetries ) on @xmath9 which are the origins of many - to - one of ( [ q11 ] ) . since ( [ q11 ] ) is a surjective representation of 3-manifold @xmath9 ( in the sense that if @xmath912 is not homeomorphic to @xmath913 then their representations ( [ q11 ] ) are not equal ) we can investigate these symmetries from the form of ( [ q11 ] ) . from ( [ q11 ] ) we see that there are three independent degrees of freedom from the three indexes of the three @xmath7-matrices : @xmath7 , @xmath914 and @xmath915 .
then we notice that there is a degenerate degree of freedom between @xmath382 and @xmath897 that @xmath906 while @xmath914 and @xmath915 act on @xmath829 and @xmath830 respectively .
further this degenerate degree of freedom is the only degenerate degree of freedom of ( [ q11 ] ) .
this degenerate degree of freedom of ( [ q11 ] ) is the source of all symmetries on @xmath9 when @xmath9 is represented by ( [ q11 ] ) since ( [ q11 ] ) is a surjective representation of @xmath9 that it contains all the topological properties of @xmath9 and that any nontrivial symmetry of @xmath9 when @xmath9 is represented by ( [ q11 ] ) needs a degenerate degree of freedom in ( [ q11 ] ) for its existence .
let us consider a symmetry from this degenerate degree of freedom .
if we reverse the orientation of @xmath897 then we have a symmetry that the obtained manifold is the same manifold as the original one . under this symmetry
we have that @xmath893 is changed to @xmath916 .
let us call this symmetry as the @xmath917-symmetry .
then there is a symmetry between @xmath382 and its mirror image @xmath918 that surgery on @xmath892 and on @xmath919 give the same 3-manifold .
this symmetry is also from the changing of @xmath893 to @xmath916 and can be regarded as a part of the whole @xmath917-symmetry .
then when @xmath382 is an amphichieral knot that @xmath920 there is a further degeneration that @xmath921 .
let us consider the @xmath917-symmetry . by the @xmath917-symmetry we have that surgery on @xmath892 and on @xmath922
give the same manifold .
let us consider the details of this symmetry step by step .
let us first consider the case that @xmath923 .
let @xmath382 be a nontrivial knot . then by the @xmath917-symmetry we have that @xmath924 where @xmath925 represents the same 3-manifold as that of @xmath926 where @xmath927 denotes the knot which is obtained from @xmath382 by reversing the orientation of @xmath382 ( we have @xmath928 since @xmath925 in this case ) .
this reversing of orientation is from the @xmath917-symmetry that @xmath646 is changed to @xmath929 .
let us then consider the structure of @xmath930 . since @xmath898 and that a general form of @xmath7-matrix is of the form @xmath931 for some integer @xmath54 and for some @xmath7-matrix @xmath914 we have that @xmath930 is of the following form : @xmath932 where we let @xmath933 and the @xmath914 of @xmath898 is replaced by @xmath934 for some integer @xmath54 which is as a new variable to be determined and we let @xmath252 be a copy of @xmath248 with reversing orientation and @xmath865 acts on @xmath904 for @xmath935 .
we notice that we now have the vector @xmath936 which is as the index vector for the @xmath914 and @xmath915 matrices where the integer @xmath54 is as a new degree of freedom when @xmath937 ( @xmath937 corresponds to a nontrivial knot ) .
this implies that the integer @xmath938 is a new degree of freedom when @xmath937 .
thus we have that from the @xmath917-symmetry a new degree of freedom is introduced and this completely eliminates the property of degenerate degree of freedom of ( [ q11 ] ) .
now since there are no more degenerate degree of freedom left in the form ( [ q13 ] ) we have that the @xmath917-symmetry is the only nontrivial symmetry which can be derived from ( [ q11 ] ) when @xmath939 where by the term nontrivial symmetry we mean a symmetry which can transform a form of ( [ q11 ] ) to another distinct form of ( [ q11 ] ) .
let us now determine the integer @xmath54 for @xmath930 for a given @xmath893 . to this end
let us consider some consequences of the @xmath940-symmetry , as follows . from the @xmath917-symmetry we have the degenerate property that @xmath941 and @xmath942 represent the same manifold .
we have @xmath943 and @xmath944 . for simplicity let us sometimes omit the factor @xmath945 .
then we have that the two distinct products @xmath946 and @xmath947 ( @xmath948 ) represent the same manifold .
this is as the degenerate property of the @xmath940-symmetry .
we have that the factor @xmath949 of @xmath946 has already been in the representation @xmath947 . thus the representation
@xmath947 contains the information of the representation @xmath946 and the information of all degenerate properties of the @xmath917-symmetry .
thus we may use the representation @xmath947 only to represent the manifold and that this representation @xmath947 contains the information of all degenerate property of the @xmath940-symmetry about the manifold .
now suppose we have another representation @xmath950 where @xmath951 and @xmath952 .
then since @xmath951 there are at least three distinct integers from the set @xmath953 and thus this is over the maximal degenerate property of the @xmath917-symmetry represented by @xmath947 which has the maximal degenerate property of the @xmath917-symmetry of allowing at most two distinct integers @xmath954 .
it then follows that the two representations @xmath947 and @xmath950 represent two nonequivalent 3-manifolds for @xmath951 .
further since the @xmath917-symmetry is the only symmetry of ( [ q11 ] ) and the matrix @xmath955 as a function of @xmath893 is a one - to - one mapping we have that the two representations @xmath947 and @xmath956 represent two nonequivalent 3-manifolds where @xmath957 when @xmath958 where @xmath893 and @xmath959 are of the same sign . in summary
we have the following theorem : the representation @xmath947 indexed by the integers @xmath960 where @xmath948 represents 3-manifolds in a one - to - one way in the sense that if @xmath951 then @xmath947 and @xmath950 represent two nonequivalent 3-manifolds where @xmath957 .
further we have that the two representations @xmath947 and @xmath956 represent two nonequivalent 3-manifolds where @xmath957 when @xmath958 where @xmath893 and @xmath959 are of the same sign .
now let us determine the property of the number @xmath54 .
we have that @xmath54 always exists since it is for the representation @xmath930 of @xmath927 .
let us consider the representation @xmath947 for @xmath961 where we consider the case that @xmath948 .
we want to find out the property of @xmath54 and the relation between @xmath54 and @xmath464 . for this @xmath54 as similar to the role of @xmath464
let us also construct a product @xmath962 .
then when @xmath963 and @xmath964 and @xmath965 by the above theorem we have that the two products @xmath947 and @xmath962 can not represent the same 3-manifold .
thus for @xmath947 and @xmath962 represent the same 3-manifold we must have either @xmath966 when @xmath964 ( this implies that @xmath967 ) or @xmath968 when @xmath969 ( this implies @xmath970 ) or @xmath971 . for the case @xmath971 we have that the product @xmath962 equals @xmath972 which represents the framed knot @xmath973 for an amphichieral knot @xmath974 with the property that @xmath975 ( and @xmath976 ) and that @xmath977 and @xmath978 represent the same manifold . for this amphichieral knot @xmath974
we have that the representation @xmath972 contains only one integer @xmath54 and thus its information is contained in @xmath947 which contains two integers @xmath8 and @xmath54 and thus @xmath972 and @xmath947 represent the same manifold where the representation @xmath947 is with the maximal non - degenerate property in the sense that it has the index vector @xmath979 where @xmath980 and @xmath981 such that no more degenerate degree of freedom left in this representation . for the case
@xmath966 we have the representation @xmath982 for @xmath983 which represents the same manifold as that of @xmath946 for @xmath984 .
for the case @xmath968 when @xmath969 and @xmath985 the representation @xmath986 for @xmath983 which represents the same manifold as that of @xmath987 for @xmath984 .
these three cases then determine the property of @xmath54 and its relation with @xmath464 .
we have that the case @xmath971 corresponds to an amphichieral knot @xmath974 . on the other hand
each amphichieral knot @xmath974 gives the representation @xmath978 of a manifold @xmath9 which is a degenerate result of the @xmath917-symmetry .
this degeneratation is as a part of a whole @xmath917-symmetry .
thus each amphichieral knot @xmath974 gives a different nontrivial homepmorphism .
thus from the above analysis there must exist a framed knot @xmath988 which gives the same manifold @xmath9 where more generally we let @xmath989 for some @xmath646 .
when @xmath923 from the above analysis we have that @xmath382 is non - amphichieral ( and thus @xmath990 ) and @xmath9 is with the representation @xmath947 where @xmath464 is the integer indexing @xmath382 .
then when @xmath991 we have that @xmath988 is represented by @xmath992 where @xmath574 is the integer indexing @xmath897 .
if @xmath993 ( or @xmath994 ) then we have the same result as the case @xmath923 that @xmath382 is non - amphichieral and @xmath9 is with the representation @xmath947 .
if @xmath995 ( or @xmath996 ) then from the above analysis we have that @xmath9 is with the representation @xmath997 . then from the above analysis we must have @xmath998 or @xmath999 ( i.e. @xmath1000 or @xmath1001 ) .
if @xmath998 then @xmath1002 and @xmath9 is with the representation @xmath1003 . comparing to the representation
@xmath972 we have that @xmath1004 .
thus if @xmath1000 the framed knot @xmath988 gives no new information to eliminate the degeneration .
thus we must have @xmath1001 ( or @xmath999 ) .
thus @xmath9 is with the representation @xmath1005 .
thus we have the following theorem : let @xmath9 be a closed ( oriented and connected ) 3-manifold which is constructed by a dehn surgery on a framed knot @xmath984 ( or on @xmath983 ) where @xmath382 is a nontrivial knot
. then @xmath9 can be uniquely represented by a representation of the form @xmath947 where @xmath464 and @xmath54 are integers and @xmath948 . further when @xmath9 is not obtained from surgery on @xmath973 where @xmath974 is an amphichieral knot we have that @xmath966 . on the other hand when @xmath9 is obtained from an amphichieral knot @xmath974 we have that the integer @xmath54 is from the representation @xmath931 of the amphichieral knot @xmath974 and the integer @xmath464 is from the representation @xmath1006 of another knot @xmath382 where @xmath9 is also obtained from surgery on @xmath988 ( or on @xmath1007 ) for some @xmath989 by the @xmath917-symmetry ( when @xmath1008 we have @xmath1001 where @xmath897 denotes the partner of @xmath382 for @xmath988 ) .
furthermore we have that the two representations @xmath947 and @xmath956 represent two nonequivalent 3-manifolds where @xmath957 when @xmath958 where @xmath893 and @xmath959 are of the same sign .
we remark that by the @xmath917-symmetry we may fix the sign of @xmath893 ( for example we may fix the sign of @xmath893 such that @xmath1009 ) to obtain the manifold @xmath9 if @xmath9 is obtained from @xmath984 or @xmath1010 .
we shall also write the above invariant @xmath947 in the following complete form : @xmath1011 let us then consider the case that @xmath9 is obtained from surgery on @xmath988 where @xmath382 is a nontrivial knot and @xmath991 is an integer which is co - prime with respect to @xmath893 . from the @xmath917-symmetry we have that @xmath988 and
@xmath1007 give the same manifold .
then we have the representations @xmath1012 and @xmath1013 for the same manifold .
we write @xmath899 where @xmath900 is the integer indexing @xmath897 and is related to the integers @xmath646 and @xmath464 where @xmath464 is the integer indexing @xmath382 . then if @xmath1014 we have that @xmath9 is a lens space and let us consider this subcase later . for
the subcase @xmath1015 suppose that there exists an amphichieral knot @xmath974 related to @xmath382 for @xmath991 as above .
then by the above theorem @xmath9 is with the representation @xmath1005 where @xmath1016 and @xmath54 is the integer indexing @xmath974 . on the other hand
suppose that there does not exist an amphichieral knot @xmath974 related to @xmath382 for @xmath991 as above .
then by following the above analysis for @xmath1017 we have that @xmath9 is simply with the representation @xmath1018 where @xmath1016 . then as similar to the above case @xmath1017 since @xmath917-symmetry is the only symmetry ( when @xmath937 ) and the matrix @xmath955 as a function of @xmath893 is a one - to - one mapping we have that two representations @xmath1005 and @xmath1019 represent two nonequivalent 3-manifolds when @xmath958 where @xmath893 and @xmath959 are of the same sign .
thus as similar to the above theorem we have the following theorem : let @xmath9 be a closed ( oriented and connected ) 3-manifold which is constructed by a dehn surgery on a framed knot @xmath988 ( or on @xmath1007 ) where @xmath382 is a nontrivial knot and @xmath1020 and @xmath9 is not a lens space . then @xmath9 can be uniquely represented by a representation of the form @xmath1021 where @xmath937 is an integer for the representation @xmath898 of @xmath382 and @xmath1022 is an integer for the representation @xmath899 of @xmath897 .
further we have that the two representations @xmath1005 and @xmath1019 represent two nonequivalent 3-manifolds when @xmath958 where @xmath893 and @xmath959 are of the same sign . from the above theorems we have the following theorem of one - to - one representation of 3-manifolds obtained from framed knots @xmath988 : let @xmath9 be a closed ( oriented and connected ) 3-manifold which is constructed by a dehn surgery on a framed knot @xmath988 where @xmath382 is a nontrivial knot and @xmath9 is not a lens space
. then we have the following one - to one representation ( or invariant ) of @xmath9 ( we call this invariant as the invariant with the maximal non - degenerate property ) : @xmath1023 where @xmath937 is the integer indexing @xmath382 and @xmath1022 such that @xmath948 is an integer related to @xmath464 and @xmath646 such that @xmath54 is either the integer indexing an amphichieral knot @xmath974 giving the same @xmath9 by dehn surgery on @xmath973 or is the integer indexing the knot @xmath897 of @xmath382 ; and we choose a convention that @xmath1009
. * proof .
* from the above two theorems we have that @xmath9 constructed by a dehn surgery on a framed knot @xmath988 can be represented by ( [ qq1 ] ) where the expression @xmath1024 absorbs all the degenerate amphichieral cases which are the only degenerate cases when @xmath937 , @xmath1015 and @xmath1009 .
thus ( [ qq1 ] ) is a one - to - one representation of @xmath9 .
this proves the theorem .
@xmath72 we remark that we shall also write the above invariant @xmath1021 in the following complete form : @xmath1025 * remark*. there exist nontrivial knots @xmath382 such that the manifold @xmath9 obtained from @xmath1026 with @xmath1014 is a lens space @xcite@xcite . before the investigation of the case of lens space
let us first consider a well - known example of the above @xmath917 symmetry from @xcite and @xcite . @xmath72
* example .
* let @xmath1027 denote the right trefoil knot @xmath1028 with framing @xmath1029 and let @xmath1030 denote the figure - eight knot @xmath1031 with framing @xmath1032 .
then as shown in @xcite we have that surgery on @xmath1027 and on @xmath1030 give the same 3-manifold @xmath9 .
then as a part of the @xmath917 symmetry ( @xmath1033 ) we have that @xmath1034 gives the same manifold @xmath9 as that of @xmath1027 where @xmath1035 denotes the left trefoil knot .
then since @xmath1031 is an amphichieral knot which is equivalent to its mirror image by the same reason we have that @xmath1036 gives the same manifold @xmath9 as that of @xmath1030 .
let us investigate this example to illustrate the above invariant ( or representation ) for this manifold @xmath9 .
we have that the generalized wilson loop for @xmath1027 is given by @xmath1037 similarly the generalized wilson loop for @xmath1034 is given by @xmath1038 then the generalized wilson loop for @xmath1030 is given by @xmath1039 where the index of @xmath1031 is @xmath385 .
then as shown in the above construction of invariant since @xmath1030 and @xmath1036 give the same manifold we have that the index number @xmath385 for @xmath1031 is absorbed to the generalized wilson loop @xmath1040 : @xmath1041 this generalized wilson loop is then by definition the unique invariant ( or representation ) for the manifold @xmath9 constructed by @xmath1027 , @xmath1042 and @xmath1034 .
we notice that for this example we have that @xmath1043 which is the index of @xmath1042 and thus the indexes of @xmath914 and @xmath915 are different .
this is the maximal non - degenerate property of the invariant ( [ ex3 ] ) in the sense that it contains the two indexes @xmath383 and @xmath385 for @xmath1028 and @xmath1044 .
@xmath72 let us then consider the case @xmath1014 for lens spaces .
we have that all lens spaces can be constructed by framed knots of the form @xmath1045 where @xmath450 denotes a trivial knot .
then we have that @xmath1014 represents a trivial knot .
then @xmath1046 represents a nontrivial knot . thus the representation ( [ qq1 ] )
can represent a lens space with linking number @xmath893 when @xmath900 is related to @xmath1047 such that @xmath1014 . thus the representation ( [ qq1 ] ) gives a representation of 3-manifolds @xmath9 including all the lens spaces .
let us then determine the number @xmath464 which can give @xmath1014 ( and thus the knot @xmath382 indexed by @xmath464 can give lens space ) .
for @xmath1048 we have that @xmath1049 gives @xmath1014 and thus gives the lens space @xmath1050 .
for @xmath1033 the case @xmath1049 is excluded since the unknot @xmath450 with framing @xmath1033 can be deleted ( and thus is not minimal where we shall give details on the concept of minimal link ) . then the @xmath1051 is represented by the constant @xmath383 ( and is not represented by @xmath1052 ) .
let us then consider @xmath1053 . from the property of lens spaces we have that @xmath1045 and
@xmath1054 give the same lens space .
this symmetry ( or homeomorphism ) can be described by the following relation @xcite : @xmath1055 for some integer @xmath0 ; and @xmath1056 where @xmath1057 for some integer @xmath1058 and we choose the region ( mod @xmath893 ) of @xmath1059 such that @xmath1060 . on the other hand
since @xmath450 is a ( trivial ) amphichieral knot from the above analysis on amphichieral knot we have that there exists a nontrivial knot @xmath382 indexed by @xmath937 ( and an integer @xmath989 ) such that @xmath1061 and @xmath1062 represent the same lens space @xmath9 ( which is constructed by @xmath1045 ) where @xmath464 is replaced by @xmath938 for some integer @xmath54 such that @xmath54 is related to @xmath1063 giving the reversing of @xmath1064 and @xmath1065 as described by ( [ lens ] ) .
thus from ( [ lens ] ) we have that @xmath1066 .
thus @xmath1067 and @xmath1068 or @xmath1069 and @xmath1070 .
let us fix the choice that @xmath1070 and @xmath1071 .
then for fixed @xmath1053 the numbers @xmath1072 with ( [ lens ] ) determine @xmath988 and that @xmath1070 mod @xmath893 . then since @xmath988 is also determined by @xmath1073 and that @xmath646 and @xmath1059 are both the longitude variables for the construction of @xmath988 we have that @xmath1074 mod @xmath893 .
thus we have the following theorem : the representation ( [ qq1 ] ) can also represent all the lens spaces when @xmath900 is related to @xmath1047 such that @xmath1014 where we let the lens space @xmath1050 be represented by ( [ qq1 ] ) with @xmath1075 and for all other 3-manifold @xmath9 we let @xmath1046 .
then for @xmath1053 we have that the lens space @xmath9 constructed by @xmath1045 is uniquely ( in the sense of mod @xmath893 ) represented by the following invariant : @xmath1076 where @xmath1077 and @xmath1078 for some integer @xmath1079 and @xmath1059 is restricted to a region mod @xmath893 such that @xmath1080 ; and the nontrivial knot @xmath382 is indexed by the integer @xmath464 where @xmath1066 for some integer @xmath54 such that @xmath1071 and @xmath1070 ; and @xmath1074 mod @xmath893 . * remark*. we do not count @xmath1051 as a lens space and that @xmath1051 is simply represented by the constant @xmath383 . @xmath72 * remark*. in the above representation we choose @xmath1081 such that @xmath1077 .
it is clear that we may choose other regions ( mod @xmath893 ) for @xmath1081 . @xmath72 * remark*. we may write ( [ qq2b ] ) in the form @xmath1082 or simply in the form @xmath1083 .
this is a degenerate form of the general form of ( [ qq1 ] ) with the degeneration that the variable @xmath1006 does not appear .
let us then consider a 3-manifold @xmath9 which is obtained from a framed link @xmath10 with the minimal number @xmath0 of component knots where @xmath589 . from the second kirby moves we may suppose that @xmath10 is in the form that the components @xmath1084 of @xmath10 do not wind each other in the form described by the second kirby moves ( we remark that the usual second kirby move is for framed links with integral framings
. in the following lemma we may generalize it to framed links with rational framings ) .
let us say that this minimal @xmath10 is in the form of maximal non - degenerate state where the degenerate property is from the winding of one component knot with the other component knot by the second kirby moves .
thus this @xmath10 has both the minimal and maximal properties as described .
then we want to find a one - to - one representation ( or invariant ) of @xmath9 from this @xmath10 . before this
let us first prove the following lemma which extends the kirby theorem : let a move generalize the usual second kirby move to framed links with rational framings .
then we have that any homeomorphism on a 3-manifold can be written as a sequence of first kirby move and this generalized second kirby move .
* proof*. let two framed knots @xmath1085 be with ( coprime ) rational framings ( for generalzing the second kirby move from integral framings to rational framings ) .
let the quantum invariant form of these two framed knots be given by : @xmath1086 for @xmath935 where @xmath865 , @xmath1087 , @xmath1088 for @xmath935 are independent ( the term @xmath1089 disappears when @xmath1090 is for a lens space ) . in terms of these two quantum invariant forms
the ( generalized ) second kirby move is described by changing @xmath1091 to the following form : @xmath1092 where the knot @xmath1093 is obtained by winding @xmath471 to @xmath472 ( by the connected sum operation ) as described by the usual second kirby move @xcite@xcite ; @xmath1094 denotes the assigned integer of @xmath1093 and @xmath1095 is the number correspoding to the number @xmath1096 ; and from the winding of @xmath471 and its partner to @xmath472 and its partner respectively we have the degeneration that in ( [ kirby2 ] ) @xmath1097 and @xmath1098 ; and the linking number between @xmath1099 and @xmath1100 is determined by the winding of @xmath471 to @xmath472 . then from ( [ kirby1 ] ) for @xmath935 we can construct ( [ kirby2 ] ) for @xmath1099 .
conversely ( by the degeneration @xmath1097 and @xmath1098 ) from ( [ kirby2 ] ) for @xmath1099 and ( [ kirby1 ] ) for @xmath1101 we can reconstruct the data for @xmath1102 and thus the quantum invariant ( [ kirby1 ] ) for @xmath1102 .
thus these two representations are equivalent .
then the degeneration @xmath1097 and @xmath1098 of this winding of generalized second kirby move gives a symmetry ( or homeomorphism ) .
conversely this winding is the only way to get a degenerate form which is equivalent to the two quantum invariant forms ( [ kirby1 ] ) .
thus this winding of generalized second kirby move is the only source for introducing symmetry relating two framed knots .
thus any homeomorphism on a 3-manifold can be written as a sequence of first kirby move and this generalized second kirby move .
this proves the lemma .
@xmath72 for simplicity
let us call this generalized second kirby move as the second kirby move .
then we want to find a one - to - one representation ( or invariant ) of @xmath9 from the given @xmath10 .
let us write @xmath850 in the form : @xmath1103 where @xmath1104 denotes a product of @xmath7-matrices acting on a subset of @xmath1105 where @xmath1106 are independent ( this is from the form of @xmath10 that the component knots @xmath638 are independent in the sense that they do not wind each other by the second kirby moves ) .
then we consider the following representation ( or invariant ) of @xmath9 : @xmath1107 where we define @xmath1108 by ( [ qq1 ] ) and they are independent .
we have the following theorem : let @xmath9 be a closed ( oriented and connected ) 3-manifold which is constructed by a dehn surgery on a framed @xmath10 with the minimal number @xmath0 of component knots ( @xmath10 has both the minimal and maximal properties )
. then we have that ( [ qq2 ] ) is a one - to - one representation ( or invariant ) of @xmath9 .
* proof*. we want to show that ( [ qq2 ] ) is a one - to - one representation ( or invariant ) of @xmath9 .
let @xmath1109 be another framed link for @xmath9 which is also with the minimal number @xmath0 ( and with the maximal property ) .
then we want to show @xmath1110 .
suppose that each component @xmath1108 of @xmath1111 does not represent a lens space .
then these components @xmath1108 are invariants of the components of @xmath10 respectively . then since the components of @xmath10 do not wind each other as described by the second kirby move we have that the components of @xmath10 are independent of each other .
thus there is no nontrivial homeomorphism changing these components @xmath1108 except those homeomorphisms involving the second kirby moves for the winding of the components of @xmath10 with each other .
then under the second kirby moves of these homeomorphisms we have that the components of @xmath10 wind each other and thus will reduce the independent degree of freedom to be less than @xmath0 .
thus to restore the degree of freedom to @xmath0 these homeomorphisms must also contain the first kirby moves of adding unknots with framing @xmath1112 . in this case these unknots can be deleted and thus @xmath10 is not minimal and this is a contradiction .
thus there is no nontrivial homeomorphism changing the components @xmath1108 of @xmath1111 except those homeomorphisms consist of only the second kirby moves for the winding of the components of @xmath10 with each other .
now suppose that @xmath1113 .
then there exists nontrivial homeomorphism of changing @xmath10 to @xmath1109 for changing the components @xmath1108 of @xmath1111 to the components of @xmath1114 .
this is impossible since there are no nontrivial homeomorphsm for changing these components @xmath1108 except those homeomorphisms consist of only the second kirby moves for the winding of the components of @xmath10 with each other .
thus @xmath1110 .
then let us suppose that there exists a component @xmath1108 of @xmath1111 representing a lens space .
then this component @xmath1090 must not be linked with the other components of @xmath10 .
suppose not .
then by the rolfsen twist on this component @xmath1090 such linking changes this component and the components linking to this component and thus the rolfsen twist is a nontrivial homeomorphism on @xmath10 and thus by the above lemma it must contains a first kirby move of adding a framed unknot @xmath1115 with framing @xmath1112 .
then this framed unknot @xmath1115 can be deleted by the first kirby move and thus @xmath10 is not minimal .
this is a contradiction
. thus @xmath10 must be in the form that it is the sum of two parts where one part is only with components which do not represent lens spaces and are of maximal nondegenerate form and the other part is formed by the components of @xmath10 representing lens spaces and each component of @xmath10 whenever representing a lens space must be unlinked with the other components of @xmath10 .
further since @xmath10 is minimal these unlinked components of @xmath10 can not be combined with each other to form another minimal representation .
thus the part of @xmath1111 formed by the components of @xmath1111 representing lens spaces is unique .
on the other hand we have also shown that the part of @xmath1111 with only components which do not represent lens spaces and are of maximal nondegenerate form is unique .
thus we have that @xmath1111 is unique and @xmath1110 .
in conclusion we have that ( [ qq2 ] ) is a one - to - one representation ( or invariant ) of @xmath9 , as was to be proved .
@xmath72 as a converse to the above theorem let us suppose that the representation ( [ qq2 ] ) uniquely represents @xmath1116 in the sense that there are no nontrivial homeomorphism transforming the @xmath0 independent components of @xmath1111 to other @xmath0 independent components of @xmath1114 where the link @xmath1109 also gives the manifold @xmath1116 .
then from the above proof we see that the link @xmath10 is a minimal ( and maximal ) link for obtaining @xmath1116 .
* remark*. let @xmath10 be a minimal ( and maximal ) framed link .
then from the above proof we have that the components of @xmath10 are independent of each other in the sense that if we transform a component framed knot of @xmath10 to an equivalent framed knot by a homeomorphism then the other components of @xmath10 are not affected by this transformation .
@xmath72 from the above theorems we then have the following classification theorem :
let @xmath9 be a closed ( oriented and connected ) 3-manifold which is not homeomorphic to @xmath1051 .
then the representation consists of ( [ qq2a ] ) , ( [ qq1 ] ) , ( [ qq2b ] ) ( or ( [ qq2c ] ) ) is a one - to - one invariant of @xmath9 .
this quantum invariant of @xmath9 has the following general expression ( which is the representation ( [ qq2 ] ) ) : @xmath1117 where @xmath10 denotes a minimal surgery link for @xmath9 and @xmath1118 is the minimal number for @xmath9 ( and @xmath10 is with the maximal property ) .
for @xmath1119 we have that the invariant for @xmath9 is @xmath383 .
let us apply the above classification of closed 3-manifolds to prove the poincar conjecture .
let @xmath9 be a closed 3-manifold obtained from surgery on a framed nontrivial knot @xmath892 which is the minimal link for @xmath9 with minimal number @xmath569 . from the above section we have the one - to - one generalized wilson loop representation @xmath1108 as invariant of @xmath9 ( from this we have that @xmath892 is a minimal link for @xmath9 with minimal number @xmath569 and when @xmath9 is a lens space we also use this invariant of @xmath9 to represent @xmath9 ) .
let us from this invariant to show that @xmath9 is non - simply connected . to this end
let us consider the fundamental group of @xmath9 .
we recall that the fundamental group of @xmath9 can be obtained from the knot group @xmath245 of @xmath382 by adding relations to the generators of @xmath245 where these additional relations are from the partner knot @xmath897 of @xmath382 . by these additional relations we have that the fundamental group of @xmath9 is formed as a subgroup of @xmath245 . as an example we have that the fundamental group of the poincare sphere @xmath9 is given by @xmath1120= x^{-2}yxz=1\}$ ] where @xmath1121 are generators of the knot group @xmath245 of the right trefoil knot @xmath382 and @xmath1122= x^{-2}yxz=1 $ ]
is the additional relation .
we have that the generators of @xmath245 are distinguished by the crossings of the knot @xmath382 .
then we have that @xmath382 is represented by the generalized wilson loop @xmath898 which is in a form that the crossings of @xmath382 have been equivalently eliminated such that @xmath382 is represented by a circle @xmath248 which winds with additional @xmath464 times by the factor @xmath1006 ( similarly we have that @xmath988 represented by ( [ qq1 ] ) is in a form that the crossings have been equivalently eliminated ) .
now let @xmath352 be a generator of @xmath245 of @xmath382 .
then we have that @xmath352 is represented as a generator of a knot group of @xmath248 in this representation @xmath898 of @xmath382 . in this representation
we have that @xmath352 is represented as a circle encircling @xmath248 .
let us write @xmath1123 for some variable integer @xmath1058 which is a form that @xmath248 winds @xmath1058 times .
thus the total winding is @xmath1124 times .
then as an equivalence we may regard @xmath248 winds one time and @xmath352 is represented as a circle encircles @xmath248 with @xmath1124 times .
now while the generators @xmath352 of @xmath245 of @xmath382 are distinguished by the crossings of @xmath382 we have that in the representation @xmath898 of @xmath382 these @xmath352 are not distinguished when they are generators for @xmath248 since @xmath248 has no crossings .
similarly for a generator @xmath1125 of the knot group of @xmath1126 represented by @xmath1127 we regard @xmath1125 as a circle encircles @xmath252 with @xmath1128 times while @xmath252 winds one time . now in the generalized wilson loop representation @xmath1129 of @xmath892 we have that all the crossings of @xmath892 are eliminated ( when @xmath892 is for a lens space we have the form ( [ qq2c ] ) that the factor @xmath1006 disappears ) . this representation is similar to the representation @xmath945 of the manifold @xmath1050 that all the crossings of @xmath248 , @xmath252 and between @xmath248 and @xmath252 are eliminated .
thus as similar to the case of the manifold @xmath1050 we have that in this representation the additional relations among the generators @xmath352 of @xmath245 for @xmath382 for defining the fundamental group @xmath1130 of @xmath9 from @xmath245 are equivalently eliminated and equivalently transformed to a relation from which the generators of @xmath382 are related to generators of @xmath1131 such that the generators of the fundamental group of @xmath9 are formed from the generators of @xmath382 .
let us determine this relation in this wilson loop representation , as follows . since @xmath1130 is a subgroup of @xmath245 we have that a generator @xmath49 of @xmath1130 in this generalized wilson loop representation is of the form that @xmath49 is a multiple product of @xmath352 that @xmath49 is a circle encircles @xmath248 with @xmath1132 times for some integer @xmath1133 .
then we have that the additional relation gives a relation of @xmath49 to the generators @xmath1125 of @xmath1131 . now in this generalized wilson loop representation we have that the only way that @xmath49 is related to the generators @xmath1125 of @xmath1131 is that @xmath49 is also a multiple product of @xmath1125 that @xmath49 is a circle encircles @xmath252 with @xmath1134 times for some integer @xmath1135 . then since @xmath938 contains the factor @xmath464 we may choose the integer @xmath1136 such that @xmath1124 is a factor of @xmath1128 that @xmath1137 .
it follows that we have the existence of @xmath49 of the form that @xmath49 is as a circle encircles @xmath248 and @xmath252 with @xmath1138 times .
now since @xmath245 is a nontrivial group with nontrivial generators @xmath352 we have that the fundamental group @xmath1130 of @xmath9 is with the existence of nontrivial generators @xmath49 which in the generalized wilson loop representation @xmath1139 are just some circles encircle @xmath248 and @xmath252 with @xmath1138 times .
this shows that @xmath9 is non simply - connected .
thus we have proved the following property p conjecture : ( property p conjecture ) let @xmath382 be a nontrivial knot .
then the 3-manifold @xmath9 obtained from dehn surgery on @xmath892 is non simply - connected . *
remark*. when @xmath892 is for a lens space and @xmath1048 we have @xmath1140
. then we have that @xmath49 is a circle encircles @xmath248 and @xmath252 for @xmath1079 times
. then any element of @xmath1130 is of the form @xmath1141 which is a circle encircles @xmath248 and @xmath252 for @xmath1142 times .
thus @xmath1130 is the group @xmath1143 of integers .
@xmath72 * remark*. when @xmath892 is for a lens space @xmath9 and @xmath1053 then the generator @xmath49 of the fundamental group @xmath1130 is a circle encircles @xmath248 and @xmath252 for @xmath1144 times for @xmath1136 .
then we have that @xmath1145 is a circle encircles @xmath248 and @xmath252 for @xmath1146 times which is of the form @xmath1147 mod @xmath893 .
thus @xmath1145 can be identified as the identity @xmath35 of a quotient group which is a cyclic group with @xmath893 elements and with @xmath49 as the generator . @xmath72
* remark*. for the poincar sphere @xmath1148 obtained from surgery on @xmath1149 we have that @xmath1150 where @xmath969 is the index for the right trefoil knot @xmath1028 .
thus the knot group of @xmath1028 and the fundamental group @xmath1151 of @xmath1148 are with the generators @xmath49 and @xmath1152 respectively and thus can be with the same generator @xmath49 . then since @xmath1151 is a proper subgroup of the knot group of @xmath1028 we have that the representation of the fundamental group @xmath1151 must be a quotient group with finite elements of the representation of the knot group of @xmath1028 which is an infinite cyclic group generated by @xmath49 .
we have that this quotient group is with element of the form @xmath1153 with windings @xmath1154 $ ] where we choose @xmath1155 for some integer variable @xmath1156 .
then when @xmath1157 we have @xmath1158 mod @xmath1159 and @xmath1160 where @xmath1159 is the number of elements of @xmath1151 .
similarly for a 3-manifold @xmath9 obtained from @xmath1161 for @xmath1053 ( the poincar sphere @xmath1148 is with @xmath1033 ) we have that @xmath1162 and thus the fundamental group @xmath1130 is a finite nontrivial group .
on the other hand when a 3-manifold @xmath9 obtained from @xmath1163 ( which is not a lens space ) and is not homeomorphic to 3-manifolds obtained from @xmath1161 for @xmath1164 we have that @xmath1165 .
thus in this case we have that the generator @xmath49 of the representation of @xmath1130 of @xmath9 is only a subgenerator of the representation of the knot group of @xmath382 .
thus we have that this representation of @xmath1130 which is generated by @xmath49 is already a proper cyclic subgroup of the representation of the knot group of @xmath382 . in this case
if there are no further conditions on this representation of @xmath1130 to be a quotient subgroup of the representation of the knot group of @xmath382 then we have that this representation of @xmath1130 is an infinite cyclic group . from this
we then have that the fundamental group @xmath1130 of @xmath9 is an infinite group .
@xmath72 now let @xmath9 be a closed 3-manifold which is classified by the following minimal invariant ( [ pr1 ] ) : @xmath1166 where @xmath10 denotes a surgery link for @xmath9 and @xmath589 is the minimal number for @xmath9 . from the above section we have that this is a one - to - one invariant for @xmath9 that the framed knot components @xmath1167 can not be eliminated .
then from this invariant ( or representation ) we have that the framed knot components @xmath1167 are independent of each other in the sense that their forms are not changed by each other though they are linked together to form the linked @xmath10 .
thus we have that the wilson loop representation of the fundamental group of @xmath9 contains the wilson loop representation of the fundamental groups of the manifolds constructed from the framed knot components @xmath1167 .
then since the wilson loop representation of the fundamental group of @xmath9 contains the wilson loop representation of the fundamental groups of the framed knot components @xmath1167 and these fundamental groups of the framed knot components @xmath1090 are nontrivial we have that the fundamental group of @xmath9 is nontrivial and @xmath9 is non - simply connected .
now let @xmath9 be a simply connected closed ( orientable and connected ) 3-manifold .
we want to show that it is homeomorphic to @xmath1051 .
let us suppose that @xmath9 is not homeomorphic to @xmath1051 .
then from the above classification theorem we have that @xmath9 is classified by a quantum invariant of the form ( [ pr2 ] ) for @xmath1118 .
thus we have that the fundamental group of @xmath9 is nontrivial and thus @xmath9 is not simply connected .
this is a contradiction
. thus @xmath9 must be homeomorphic to @xmath1051 , as was to be proved .
this proves the following poincar conjecture : let @xmath9 be a closed ( orientable and connected ) and simply connected 3-manifold
. then @xmath9 is homeomorphic to the 3-sphere @xmath1051 .
in this paper from a quantum gauge model we derive a conformal field theory structure from which we derive a knot invariant related to the homfly polynomial .
the relation between these two invariants is that both the homfly polynomial and this knot invariant can be derived by using two knizhnik - zamolodchikov ( kz ) equations which are dual to each other and are derived from the quantum gauge model . in this derivation
an important concept called the wilson lines and the generalized wilson loops is introduced such that each knot diagram is represented by a generalized wilson loop where the upper crossing , zero crossing and undercrossing of two curves can be represented by the orderings of two wilson lines represent these two curves .
we show that this invariant can classify knots by showing that the generalized wilson loop of a knot faithfully represents the knot in the sense that if two knot diagrams have the same generalized wilson loop then these two knot diagrams must be equivalent .
this invariant is in terms of the monodromy @xmath7 of the two knizhnik - zamolochikov equations . in the case of knots
this invariant can be written in the form @xmath478 . from this invariant
we may classify knots with the integer @xmath8 . a classification table of knots
can then be formed where prime knots are classified with odd prime numbers @xmath8 and non - prime knots are classified with non - prime numbers @xmath8 .
then from the quantum link invariant we can construct quantum invariant of 3-manifolds .
we first construct quantum invariant of closed three - manifolds obtained by dehn surgery on framed knots .
we then introduce the concept of minimal link to construct quantum invariant of closed three - manifolds obtained by dehn surgery on framed links . then by using the lickorish - wallace theorem
we show that this quantum invariant of 3-manifolds gives a one - to - one classification of closed 3-manifolds . from this classification of closed 3-manifolds
we can then prove the poincar conjecture . | a new quantum gauge model is proposed . from this quantum gauge model
we derive a quantum invariant of 3-manifolds .
we show that this invariant gives a classification of closed ( orientable and connected ) 3-manifolds . from this classification
we prove the poincar conjecture .
* mathematics subject classification : * 57m27 , 51p05 , 81t10 , 81t40 . |
synthetic copolymers have numerous industrial applications due to their rich phase diagram and ability to tailor particular physical properties by a choice of the monomers used and the order in which they are polymerised .
generally , a synthesis would produce a quasi random arrangement of the monomers , which can be characterised by the probability distribution of the composition variables @xmath1 , where one agreed to assign @xmath2 if the monomer @xmath3 is of type @xmath4 and @xmath5 if the monomer @xmath3 is of type @xmath6 .
the overall composition is then given by the ratio of the total number of @xmath4 and @xmath6 species , and it would typically be fixed and determined by the ratio of reactants concentrations during polymerisation , provided that further separation techniques have allowed one to ensure a good degree of monodispersity in the degree of polymerisation @xmath7 . likewise , by special procedures one can also prepare well controlled primary sequences @xmath1 corresponding to di blocks , tri blocks , and more generally shorter alternating blocks of @xmath4 and @xmath6 units .
some of perhaps most well studied examples of synthetic water soluble polymers of these types would include pnipam , peo - ppo systems and their various modifications @xcite .
most important properties of monomers from the point of view of molecular modelling are their degree of hydrophobicity and effective charges , and these along with the primary sequence would determine the conformations of the copolymers in solution at given thermodynamic conditions . both concepts
, however , involve rather complex elementary inter atomic interactions : degree of hydrophobicity results from a competition of the van der waals , steric , hydrogen bonding and solvent entropic effects , while the effective charges depend on the electronic properties of the polymer , solvent dielectric permitivity and degree of screening by the counter
ions present in solution . while full
scale atomistic simulations are rather standard for concrete polymers in solutions nowadays , their ability to study thermodynamic universal properties of whole classes of macromolecules are rather limited due to unbearable computational expenses involved . for these purposes
a number of simplified coarse
grained copolymer models , which phenomenologically capture the effects of hydrophobicity and charges , have been proposed and studied both analytically @xcite and by monte carlo simulation ( see e.g. @xcite and references therein ) , often on lattice , which further significantly reduces computational expenses .
a great deal of related research has also been undertaken in relation to the problem of protein folding in the context of random heteropolymer models of proteins @xcite .
although natural proteins are composed of 20 types of amino acid residues with rather refined properties , in a simplified model one can roughly divide all amino acid residues into polar ( hydrophilic ) and apolar ( hydrophobic ) , as well as into effectively charged and neutral .
this broad division produces 2-letter models of proteins , and since the primary sequence of each protein is unique , applied to a whole ensemble of sequences this leads to random heteropolymer models of proteins . depending on which particular feature one is trying to describe thus there are models of amphiphilic @xcite and charged copolymers @xcite after solvent molecules have been effectively excluded from the consideration by integration ( see refs .
@xcite for details ) .
the first model has a linear term in composition variables , @xmath8 , in the hamiltonian , so that hydrophobic units @xmath4 effectively attract each other ( and thus dislike solvent contacts ) , hydrophilic @xmath6 units effectively repel each other ( and thus like solvent contacts ) , whereas hydrophobic - hydrophilic @xmath9 direct pair
wise interactions are zero . such model will be referred to as hydrophobic hydrophilic or hh model . as a simple variation of this we can also introduce hydrophobic - ideal or hi model , in which @xmath6 species have zero repulsion from each other and thus effectively behave as ideal . the model with a quadratic term in composition variables , @xmath10 , in the hamiltonian
is often called the random ` charge ' model . in this context
one can introduce the true charge or briefly ca model , in which likewise charges repel and opposite charges attract each other , or the model in which the quadratic term appears in the hamiltonian with the opposite sign , so that likewise charges attract and opposite ones repel each other , which is naturally to call the anti charge or ac model . although the latter model does not describe true atomistic charges , such terms nevertheless appear to reproduce complex non local effective monomer solvent interactions arising after coarse graining of models with complex intra
molecular potentials .
namely , the ac model was most popular in literature in analytical and lattice monte carlo studies of random heteropolymer models of proteins @xcite .
as we shall see later it indeed has the most peculiar properties , and although its applicability to synthetic copolymers is least justified , no single model is good enough for describing the complex behaviour of proteins . in this work
we would be interested to extensively investigate the energy probability distribution functions @xmath11 in the hh , hi , ca and ac models for a large set of different heteropolymers sequences at low temperatures after transitions to compacted states , whether crystalline or glassy , as well as to try to relate the observed distributions to copolymer conformations , by which we understand the arrangement of monomers in relation to each other in space .
the energy probability distribution is formally defined as , @xmath12 where @xmath13 refers to microstates of the system , and this is linearly related to the density of states times an explicit exponential factor and a constant @xmath14 , @xmath15 we note that although in ref .
@xcite monte carlo histogram technique was successfully applied to a 27-mer on a cubic lattice without enumerating all compact 3x3x3 cube states , here we shall be able to investigate many copolymer sequences of a longer chain on a large lattice with a good statistical sampling able to distinguish universal patterns of behaviour in the four aforementioned copolymer models directly at low temperatures without using extrapolation by means of eq .
( [ eq : ne ] ) .
this was previously problematic as a much larger sampling is required at temperatures below the transitions to compacted states .
we adopt the metropolis technique in the lattice model of our ref . @xcite .
this model , apart from the connectivity and excluded volume constraints , includes short ranged pair
wise interactions between lattice sites . in case of copolymers
one has to distinguish two types of monomers _ a _ and _ b _ and ascribe to them three values of flory interaction parameters as follows @xcite , @xmath16 as in ref .
@xcite we shall use the parametrisation such that @xmath17 , @xmath18 and @xmath19 thus , all interaction parameters can be summarised as in table [ tab ] .
we may note that the hydrophobic hydrophilic and hydrophobic ideal models use the same formal expression , but different values of @xmath20 and @xmath21 .
in our study we have considered copolymers of length @xmath22 units with equal number of @xmath4 and @xmath6 monomers in each sequence on a large enough cubic lattice of linear dimension @xmath23 in order to make the influence of boundary conditions negligible . for each of four models we have analysed a number of periodic and 1000 randomly generated sequences .
every initial random walk conformation was first subjected to a slow quasistatic equilibration corresponding to gradually decreasing temperatures until the values of @xmath24 presented in table [ tab ] have been reached .
this procedure is similar to ` simulated annealing ' and permits to overcome excessive trapping of the system in metastable states .
moreover , to be able to deal with the issue of ` non ergodicity ' this cooling was applied to a fairly large ensemble ( typically over 200 ) of different initial conditions . to obtain probability distributions of various observables , such as e.g. energy
, we then performed hundreds of measurements separated by large number of monte carlo sweeps ( i.e. typically hundred of thousands of attempted monte carlo moves ) , resulting in the total number of independent statistical measurements equal to 20000 in each case .
finally , normalised energy histograms provide monte carlo approximations to the exact energy probability distribution functions @xmath25 defined by eq .
( [ eq : pe ] ) . due to the enormous amount of data obtained and obvious constraints of the paper size
we shall only discuss here most interesting and representative results , while the whole data set is available online @xcite for an interested reader . a typical energy distribution function of hi model shown in fig .
[ fig:1 ] has a bell shape .
note that low energy states contribute little to this function because the density of states @xmath26 is rapidly decreasing with decreasing energy @xmath27 .
the general outlook of @xmath25 changes little for different sequences , with only mean value @xmath28 and width @xmath29 of the distribution being sensitive on the sequence . for most sequences
the energy distribution function in the hh model is similar to that of the hi model .
however , a few percents of sequences in the hh model also possess a second peak in @xmath25 as shown in fig .
[ fig:2 ] .
the secondary peak at higher energies is considerably lower and could be simply understood by visualising typical copolymer conformations corresponding to a given energy distribution .
first of all , a snapshot in fig .
[ fig:3]a shows a typical 3-d polymer shape for a single peak situation .
a clear feature of both the hi and hh models is the micro phase separation ( mps ) of the globule onto a hydrophobic ( black ) core and a hydrophilic ( gray ) shell as in a micelle . as for the case of a sequence with a two
peak @xmath25 function , this would correspond to a dimer ( or a multi
mer for longer chains ) of sub - globules connected by a predominantly hydrophilic bridge .
obviously , such a situation can be realised only for sequences which do have a fragment in their primary sequence with a predominantly hydrophilic units .
the energy distributions in the ca model also have a trivial single peak shape as in fig . [ fig:1 ] .
however , the polymer conformations are rather different in this case . thus , in fig .
[ fig:4 ] one can see that although the globules are still compact , the internal arrangement of the monomers within are rather peculiar and have an alternating rather than a core and shell structure .
this arrangement obviously tends to maximise the number of @xmath30 contacts favourable in this model . figs .
[ fig:4]a and b corresponding to a random and a long blocks sequences respectively show , however , that there is no tendency of forming better structured conformations ( in the sense of maximising the mps between @xmath4 and @xmath6 units ) for periodic sequences with increasing block length as was the case in the hh and hi models .
most interesting energy distributions of the four considered models occur in the ac model .
although most sequences here would still have a single shape distribution as in the other models , there are a few percents of sequences in the ac model which have energy histogram as shown in fig .
[ fig:5 ] .
these sequences possess a highly populated visible lowest energy state , which is often separated by a gap from the rest of the spectrum .
this may be rationalised as occurring due to much a higher density of states @xmath26 at low energies .
this unusual behaviour of @xmath25 also corresponds to interesting conformations of the copolymer . in figs .
[ fig:6]a and [ fig:6]b these are shown for examples of periodic sequences of different block lengths . for long blocks sequence in fig .
[ fig:4]a there is a clear dumbbell shape with @xmath4 and @xmath6 units in their separate homopolymer like subglobules avoiding unfavourable contacts with each other . for shorter blocks sequences as in fig .
[ fig:4]b this separation of @xmath4 and @xmath6 units can not be as perfectly realised due to the chain connectivity constraints , thus producing merely a multi
domain globule of homopolymer like clusters .
also , the increase of the block length shifts the @xmath25 function towards lower energies due to a better optimisation of competing pair wise interactions .
as we have seen here the energy probability distributions are bell shaped and shifted towards low energies in most cases .
the second peak in the hh model occurring for some sequences , especially those having a long hydrophilic segments in the chain , corresponds to formation of di mers and multi mers .
naturally , the hi model has similar features to hh model , although it has a lower tendency for forming multi - mers .
speaking of the conformations , only the hh and hi models produce good micro
phase separation ( mps ) with a predominantly hydrophobic core and hydrophilic surface .
the degree of mps increases with increasing a characteristic repeating block in a sequence .
conformations in ca model do not have a mps structure and are rather complex amorphous ( glassy ) globules .
the characteristic block length has little effect on the shape and energy in this model .
the ac model has most interesting and most sequence
dependent energy distributions of all .
it has a well populated lowest energy state separated by a gap from higher energy states for certain sequences , which perhaps can be viewed as _ good folding sequences _ according to ref .
this shape is generally believed to provide good accessibility and stability of the lowest energy state identified as an analog of the native state in proteins .
we may note that this result agrees with fig . 2 in ref .
@xcite , in which the ac model was in fact studied at somewhat different values of parameters as compared to table [ tab ] .
however , the ac model exhibits phase separation of @xmath4 and @xmath6 species similar to two immiscible liquids under connectivity constraints , which is rather different from @xmath4-surface and @xmath6-core mps in hh model .
increasing typical block length here leads to dumbbell structures rather than micelles , which also has an effect of lowering the energy considerably .
short block sequences produce multi
domain globules of identical species .
thus , while ac model has an energy histogram which bears resemblance to that of a protein in the native state , it has conformations which have little to do with typical protein shapes , especially in terms of mps .
the hh model , on the other hand , has conformations similar to those of proteins , but rather trivial energy histogram with easily accessible low energy states and no barrier or special pathway to these states from the coil .
thus , unfortunately none of the commonly used coarse - grained lattice models bears any direct relevance for modelling proteins , at least at the level of 2-letter models .
we have presented results for four distinct models of copolymers on lattice obtained from monte carlo simulations of a large ensemble of 20000 independent initial conditions for 1000 randomly generated and a number of regular sequences with equal ratio of @xmath4 and @xmath6 monomers .
features of energy distribution functions ( histograms ) and corresponding polymer shapes ( conformations ) are discussed in the hydrophobic - hydrophilic ( hh ) , hydrophobic - ideal ( hi ) , charge ( ca ) and anti - charge ( ac ) models of copolymers at temperatures much below transitions to compacted states . most sequences in all of models
have a single bell shaped energy probability distributions . however , some sequences in the hh model have two peaks distributions
. moreover , a number of sequences in ac model have more than two peaks and often a gap between the lowest energy and higher energy states .
conformations and micro phase separation of @xmath4 and @xmath6 species corresponding to distinct distributions in these models are also analysed and these results may be of interest for a range of synthetic polymers in solution .
finally , from our analysis it appears that while separate features expected from proteins can be represented by the hh and ac models of copolymers , neither model possesses all of these features together .
thus , a proper continuous space treatment with the full account of coulomb , van der waals , bonded and other specific interactions may be necessary to capture even in a minimal way the folding behaviour of proteins .
the authors acknowledge most interesting discussions with professors f. ganazzoli , t. garel , h. orland , and p.g
. wolynes at different stages of the current research .
this work was supported by grants sc/99/186 and bc/2001/034 from enterprise ireland .
j. lindman , h. kronberg , surfactants and polymers in aqueous solution , + wiley ( 1998 ) chichester . c. d. sfatos , a. m. gutin , e. i. shakhnovich , phys .
e 48 ( 1993 ) 465 ; + phys .
e 51 ( 1995 ) 4727 .
pande , a.yu .
grosberg , t. tanaka , j. chem .
( 1994 ) 8246 ; + phys .
e 51 ( 1995 ) 3381 .
d. thirumalai , v. ashwin , j.k .
bhattacharjee , phys .
77 ( 1996 ) 5385 .
e.g. timoshenko , yu.a .
kuznetsov , k.a .
dawson , phys .
e 54 ( 1996 ) 4071 .
kuznetsov , e.g. timoshenko , k.a .
dawson , j. chem .
103 ( 1995 ) 4807 .
t. garel , h. orland , europhys .
lett . 6 ( 1988 ) 597 ; + t. garel , l. leibler , h. orland , j. phys .
( france ) ii 4 ( 1994 ) 2139 .
e.g. timoshenko , yu.a .
kuznetsov , j. chem .
112 ( 2000 ) 8163 .
kuznetsov , e.g. timoshenko , il nuovo cimento .
20d ( 1998 ) 2359 .
f. ganazzoli , yu.a .
kuznetsov , e.g. timoshenko , macromol .
theory simul . 10
( 2001 ) 325 .
socci , j.n .
onuchic , j. chem .
103 ( 1995 ) 4732 .
e.g. timoshenko , yu.a .
kuznetsov , r.n .
+ url : http://darkstar.ucd.ie/data/idmrcs4.zip .
.values of the flory interaction constants for four different models of copolymers on lattice . [ cols="<,^,^,^,^,^",options="header " , ] | based on large scale monte carlo simulations on lattice the energy probability distribution functions are investigated for a large set of primary sequences in distinct models of copolymers at low temperatures below transitions to compacted states .
amphiphilic copolymers with hydrophobic and hydrophilic units are found to produce a single or double peak energy distributions corresponding to mono- or multi meric micellar conformations .
however , copolymers with short ranged random ` charge ' interactions in some cases are found to produce energy distribution functions with a well pronounced lowest energy state and a gap separating it from the rest of the spectrum .
these , however have rather peculiar conformations corresponding to effectively immiscible domains comprised from monomers of likewise species .
relevance of these observations for coarse grained models for protein folding is discussed .
@xmath0 + * pacs numbers : 36.20.-r , 87.15.-v * # 1#2i_#1#2 -3 mm |
to maintain the desired luminosity of the nlc , the focusing components on the main linac must be kept at a few nanometers above a few hz . these components can be affected by far - field ( natural ) and near - field ( man - made ) vibration sources .
this paper is concerned only with near - field sources ( e.g. mechanical and electrical equipment , rf generating equipment , etc ) .
these sources are mainly located either in the support tunnel , fig.[2tunnel ] , or far away ( @xmath0 100 m ) from the beam tunnel .
the characterization of near - field vibration sources and its effects on the main linac components is part of an ongoing r&d program at nlc that is presented in this paper .
the first part of the paper will present the influence of the vibration induced by rf power generating elements and by the rf itself . the second part will deal with the transfer of vibration from surface to the tunnel invert .
[ 2tunnel ]
at first , we will focus on the vibration contribution of high power generating rf components .
the study was carried out near the supporting structure of the 8 pack project @xcite .
the 8 pack project is the test bench of klystrons and their modulator to produce high power rf for the next linear collider test accelerator ( nlcta ) .
one of the geophone sensors was placed at the base of the modulator at 76 cm above the concrete floor and the second one was located on the concrete floor at @xmath11.2 m away from the base of the modulator support , or at its base ( 0 m ) .
the signals of the mark-4 geophones were measured simultaneously .
we performed tests in two different conditions of the modulator . in the first case , the water cooled modulator was under a 20 kv voltage and was running at 30 hz or 10 hz ( at this voltage the klystrons was not conducting , thus did not deliver rf power ) . in the second case the modulator was running at 10 hz and
was under 400 kv ( the klystron was delivering @xmath140 mw of rf power at 1.6 @xmath2s ) .
fig.[speedtimesensor ] shows the response of the sensors vs times for the modulator under a 20 kv . in this case
the modulator was running at different frequencies .
the seismometer located at the base of the modulator clearly shows vibrations due to modulator pulsed operation , while the geophone located on the concrete floor does not indicate any change due to the modulator running conditions .
[ speedtimesensor ] fourier analysis also does not reveal any additional noises due to modulator on the floor .
fig.[displc8packjan ] and fig.[displc8packfeb ] show the average integrated displacement ( aid ) of the two seismometers .
the blue lines are the response of the sensor placed at the base of the modulator , where the difference between the modulator on and off cases is clearly seen .
the green lines are the response of the sensor placed on the concrete floor either at @xmath11.2 m or at the base of the stand ( 0 m ) , and the on and off cases overlap .
interesting that , when the modulator is under 400 kv and delivering power to the klystron , its vibration is slightly less than for a 20 kv running modulator .
one can also note that background noise was different in the conditions of fig.[displc8packjan ] and fig.[displc8packfeb ] ( see green curves below 60 hz ) which was due to different level of activity of the construction crew working at nlcta .
the main conclusion of the modulator vibration study is that the transmission of vibration from the modulator to the concrete floor is not significant , and , on the level of background noise at nlcta , not noticeable .
[ displc8packjan ] [ displc8packfeb ] we also studied vibrations of accelerating structures due to cooling water ( reported earlier , see @xcite ) , and due to rf pulse ( presented below ) . when the klystrons are delivering power , the rf heating produces acoustical vibration in the accelerating structure .
possible vibration produced by the rf were measured by 3 piezo - accelerometers placed on the accelerating travelling wave structure bolted on its support or girder , placed on the girder , and placed on the waveguide .
the x - rays coming out from the structure when filled by the rf did not affect the measurements of our accelerometers . [ rfinnlcta ] these measurements , fig.[rfinnlcta ] , show that feeding a structure ( 60 cm long travelling wave structure h60vg3r ) , via its waveguide , with 100 mw at 400 ns of rf power at 60 hz ( corresponded to about 70 mv / m accelerating gradient ) does not lead to any significant increase of vibration in comparison to vibrations produced by cooling water .
the water induced vibrations dominate , but they are tolerable , since , with an appropriate design , the vibration transmission to the linac quadrupoles is rather small @xcite .
vibration of the loosely supported rf waveguide is higher than of the rf structure , but apparently this does not significantly increase vibration of either structure or the quadrupoles .
one of the questions to be addressed is whether there is a significant difference in the vibration attenuation characteristics between a tunnel bored at great depth ( in bedrock ) or excavated in cut - and - cover construction at lesser depth , with regard to vibrations generated at the surface .
a study was carried out at slac @xcite , representing the vibration attenuation characteristics of the soil , assuming the slac beam line housing to be representative of a cut - and - cover construction .
the second vibration measurement study is currently underway in the red line tunnels in los angeles .
this study will establish the vibration characteristics between two tunnels , along the tunnel as well as from surface to the tunnel .
the following is a brief presentation of data from the study carried out at slac .
[ relativeloc ] fig.[relativeloc ] shows the relative locations of sources and the receivers .
the receivers locations were on the floor along the centerline of the beam .
the sources were on the ground surface .
drive location s1 was at the same approximate elevation as of the klystron gallery floor , which lies about 36 ft ( 11 m ) above the elevation of the receiver locations .
drive location s2 was about 10 ft ( 3 m ) uphill from s1 .
[ hammerblow ] fig.[hammerblow ] shows two spectra .
one is the ambient measured at the receiver location r1 .
the other is a fast fourier transform of the response to several hammer blows .
several observations may be made .
the peaks at 60 hz and 120 hz are electronic artefacts .
the peaks centered at @xmath18.75 hz and at @xmath114.25 hz in both ambient and response have a nearly identical amplitudes .
the peak at 18.75 hz has a slightly different amplitude . at these frequencies ,
the `` response '' ( red line ) is being governed more by ambient than by the input force , so the transfer functions derived from the `` response '' will be invalid .
the `` ambient '' ( blue line ) at frequencies less than 7 hz ( in this case ) lies above the `` response '' .
this suggests that there is some variability to the ambient environment at these low frequencies , and these will degrade the accuracy of transmission function at lower frequencies .
[ groundtransmis ] fig.[groundtransmis ] shows a smoothed spectra which were calculated by log averaging the amplitudes over a 10 hz interval centered on the plotted point .
it provides the transfer function ( showing change of amplitude ) measured at r1 to r5 using s1 as a drive point , fig.[relativeloc ] .
.attenuation factor , refering to fig.[relativeloc ] .
[ cols="^,^,^,^,^ " , ] the table [ tblea ] summarizes the apparent attenuation at several frequencies at which mechanical equipment usually operates .
the figures in table [ tblea ] represent the attenuation factor a for a vibration with its source near s1 propagating along the same path .
let us give an example how such data can be used .
suppose a pump is installed at s1 , and it produces vibrations at 30 hz of amplitude x. the amplitude at 30 hz that we would measure at r5 would be the greater of either ambient or 0.0009x .
suppose we take the ambient measured shown for the tunnel in ( fig.[hammerblow ] ) as representative the tunnel in general .
the amplitude at 30 hz is about 1.5 @xmath2 m/s . if we were to place a pump at s1 and be sure to avoid having its
vibrations exceed ambient at r5 , we would need to impose a limit on the resulting vibration at s1 of 1.5/0.0009=1.7 mm / s .
it has been shown that the vibration transmitted by the rf generating equipment to the floor is insignificant .
hence , klystrons and or modulators running in the support tunnel of the nlc should not effect alignment of the linac . vibration contribution of an rf pulse to an accelerating structure has also been found negligible relative to water - cooling .
thus , it leaves electrical and mechanical rotating equipment as possibly a dominating source of vibration .
the attenuation factors presented in the paper can be used in planning stage of the nlc project for specifying and locating the mechanical rotating equipment , as well as to assess their vibration effects on the focusing components on the main linac and provide a means for establishing the vibration budgeting scheme for the project . | the vibration stability requirements for the next linear collider ( nlc ) are far more stringent than for the previous generation of colliders . to meet these goals
, it is imperative that the effects of vibration on nlc linac components from near - field sources ( e.g. compressors , high vacuum equipment , klystrons , modulators , pumps , fans , etc ) be well understood . the civil construction method , whether cut - and - cover or parallel bored tunnels , can determine the proximity and possible isolation of noise sources
this paper presents a brief summary and analysis of recently completed and planned studies for characterization of near - field vibration sources under either construction method .
the results of in - situ vibration measurements will also be included . |
the abundance ratios of chemical elements that are produced on different time scales by stars of different masses are essential probes of the history of chemical enrichment in different types of stellar populations .
they tell which stars and in which proportions have contributed to the chemical evolution of a given system at any time .
after tinsley ( 1979 ) , it has become customary to explain the trend of oxygen relative to iron as due to the different roles played by type ia supernovae ( sneia ) and type ii supernovae ( sneii ) in the chemical enrichment of galaxies ( see also greggio & renzini 1983 ; matteucci & greggio 1986 ) .
this interpretation is known as the ` time - delay model ' .
oxygen is produced mostly by core - collapse sne on very short time scales , of the order of a few myr to tenths of myr .
the bulk of iron , instead , comes later from snia explosions , on time scales ranging from 30 myr to a hubble time .
hence , the [ o / fe ] ratio in a given system is high as long as sneii dominate the interstellar medium ( ism ) pollution , then decreases to solar and subsolar values when sneia become the major contributors to the production of fe .
the maximum snia rate depends on the adopted progenitor model , as well as on the assumed star formation history ( matteucci & recchi 2001 ) ; thus , the measurements of [ o / fe ] are a powerful diagnostics of the star formation history in galaxies . of particular interest are those elements whose yields depend on the metallicity of the parent stars ; manganese ( mn ) is one of them .
mcwilliam , rich & smecker - hane ( 2003 ) confronted the [ mn / fe ] versus [ fe / h ] relation in the galactic bulge , in the solar neighbourhood and in the sagittarius dwarf spheroidal galaxy ( sgr dsph ) .
they suggested that the mn yields from both sneia and sneii are metallicity dependent .
their proposals were in agreement with extant nucleosynthesis calculations for massive stars ( e.g. arnett 1971 ; woosley & weaver 1995 ) , but there was not modelling of explosive nucleosynthesis in sneia supporting their findings .
shetrone et al .
( 2003 ) measured the mn abundances of a dozen individual red giant stars in the sculptor , fornax , carina and leo i dsphs .
they also conclude that snia contributions to the synthesis of mn must be metallicity dependent , with very little mn produced until [ fe / h ] = @xmath11 . later on , ohkubo et al .
( 2006 ) showed that , indeed , [ mn / fe ] and [ ni / fe ] in the ejecta of sneia depend on metallicity of snia progenitors .
furthermore , cescutti et al . (
2008 ) demonstrated , by means of self - consistent chemical evolution models , that the run of [ mn / fe ] with [ fe / h ] in the three independent stellar systems
the galactic bulge , the solar neighbourhood and the sgr dsph can be understood _ only in terms of a metallicity - dependent yield of mn _ from sneia .
the time - delay model alone is insufficient to explain the behaviour of [ mn / fe ] versus [ fe / h ] in the three systems .
cescutti et al .
( 2008 ) propose that the yield of mn in sneia increases with the metallicity of the progenitors , @xmath2 .
this is in agreement with the tight correlation between the mn - to - cr mass ratio in the ejecta of sneia and the metallicity of the progenitor found by badenes , bravo & hughes ( 2008 ) , @xmath3 = 5.3 @xmath4 @xmath5 .
more recently , mn abundances have been studied for the first time on a significant metallicity range in the peculiar globular cluster @xmath0centauri ( cunha et al .
2010 ; pancino et al .
in the metal - poor regime , cunha et al . ( 2010 ) find that the lte values of [ mn / fe ] in @xmath0cen stars overlap those of their solar neighbourhood analogues .
however , at variance with the solar neighbourhood trend , [ mn / fe ] declines in more metal - rich stars ( see also pancino et al .
non - lte calculations confirm the conclusion of a well distinct pattern of [ mn / fe ] versus [ fe / h ] in @xmath0cen ( cunha et al .
it is worth noticing that , in the metallicity range @xmath12.7 @xmath6 [ fe / h ] @xmath70.7 , all other galactic globulars have mn abundances equivalent to those of halo field stars ( sobeck et al .
2006 ) .
although historically classified as a globular cluster , @xmath0cen is possibly the naked nucleus of a small galactic satellite captured by the milky way many gyr ago ( e.g. freeman 1993 ; bekki & freeman 2003 ) . as such , it likely suffered a complex chemical enrichment history , marked by the occurrence of strong differential galactic winds ( romano et al .
2007 , 2010a ) .
the challenging bet is : can the low values of [ mn / fe ] observed in the more metal - rich stars of @xmath0cen be explained in the context of its peculiar evolutive history ? in this paper , we contrast the evolution of mn in the milky way with that in @xmath0cen . the goal of this work is twofold .
first , we want to add the analysis of another distinct stellar population to previous theoretical study of the evolution of mn in different environments by cescutti et al .
second , we want to get better insight into the mechanisms of formation and evolution of @xmath0cen . the predictions of our models for the milky way and for @xmath0cen are compared to both lte and non - lte mn abundances in the two systems .
the paper is organized as follows : in sect . 2
, we briefly review the relevant observational data . in sect . 3 , we describe the adopted chemical evolution models . in sect . 4
, we discuss the model results . in sect . 5
, we draw our conclusions .
for the purpose of comparison with the results of our models , we use lte mni abundances of galactic stars by cayrel et al .
( 2004 ) , gratton et al .
( 2003 ) , reddy et al .
( 2003 ) , reddy , lambert & allende prieto ( 2006 ) and feltzing , fohlman & bensby ( 2007 ) .
these studies are selected to cover the whole metallicity range of solar neighbourhood stars .
other recent measurements of mn in galactic disc and halo stars can be found in nissen et al .
( 2000 ) , prochaska & mcwilliam ( 2000 ) , norris , ryan & beers ( 2001 ) , carretta et al .
( 2002 ) , bai et al .
( 2004 ) and lai et al .
( 2008 ) . for @xmath0cen
, we use lte data from cunha et al .
( 2010 , 10 stars ) and pancino et al . ( 2011 , 6 stars ) . while cunha et al . ( 2010 ) analysed mn lines around 6000 , pancino et al . (
2011 ) analysed mn lines around 4000 .
this could explain the marginal disagreement between the two data sets and lead to an artificial increase of the scatter in the data when the two data sets are plotted together ( fig .
[ fig : lte ] , right panel ) .
cohen ( 1981 ) and gratton ( 1982 ) also published mn abundances of @xmath0cen stars .
however , their determinations do not take hyperfine splitting of spectral lines into account and are , thus , not considered here .
despite the adoption of different methods and assumptions to determine the mn abundances , all the studies mentioned above conclude that mn is deficient compared to iron ( [ mn / fe ] @xmath80.5 ) in metal - poor stars and that [ mn / fe ] increases with increasing [ fe / h ] from the halo to the thin disc populations .
the stars in @xmath0cen present a trend at odds with that of their solar neighbourhood counterparts , i.e. a [ mn / fe ] ratio declining with increasing [ fe / h ] ( cunha et al . 2010 ) .
notice , however , that this statement rests with the determination of mn in only two stars . in the lte approximation ,
the abundances based on mni lines could be underestimated by as much as 0.4 dex at low metallicities , with non - lte effects being less pronounced in high - metallicity stars ( bergemann & gehren 2008 ) . up to now ,
non - lte mn abundances have been computed only for a few stars in the milky way ( bergemann & gehren 2008 ) and in @xmath0cen ( cunha et al .
when the non - lte corrections are applied to solar neighbourhood stars , a shallower rise is found from slightly subsolar values in the halo ( [ mn / fe ] @xmath80.1 ) to solar ratios in the thin disc .
the non - lte abundance analysis of @xmath0cen stars fully confirms the odd behaviour of decreasing mn with increasing [ fe / h ] depicted by lte studies ( see discussion in cunha et al .
2010 ) , but once again the result bases on the analysis of only a handful of stars .
we discuss further the non - lte mn abundances in comparison to our model results in sect .
the chemical evolution model for the solar neighbourhood is basically the ` two - infall model ' of chiappini , matteucci & gratton ( 1997 ) , except for the adopted stellar lifetimes , nucleosynthesis and initial mass function ( imf ; see romano et al .
2005 , 2010b ) .
in particular , the kroupa , tout & gilmore ( 1993 ) imf is assumed instead of the scalo ( 1986 ) one .
this makes the predicted present - day snia - to - snii rate ratio to agree with the observations ( romano et al .
2005 , their table 4 ) and allows to better explain the evolution of deuterium in the galaxy ( romano 2010 ) .
the adopted model for @xmath0cen assumes that this cluster was once located at the centre of a more massive system , that evolved in isolation before being accreted and almost totally disrupted by the interaction with the milky way . because of the shallow potential well , the chemical evolution of the original system and of the embedded proto - cluster turns out to be significantly affected by galactic outflows triggered by multiple sn explosions ( romano et al . 2007 ) .
these outflows deprive the original system of a large fraction of its metals , thus allowing the observed na - o anticorrelation and the extreme level of he enhancement to set up in a ( minor ) fraction of the hosted stars ( romano et al .
2010a ) . for details about the basic assumptions and equations of the models , we refer the reader to the papers quoted above .
as for the adopted nucleosynthesis prescriptions , they can be found below .
as already mentioned , fe and mn are produced by both sneii and sneia , though in different proportions . in our computations , we explore the consequences of adopting six different prescriptions about the synthesis of fe and mn in stars : 1 .
nuc 1 : we use the metal - dependent yields of woosley & weaver ( 1995 ) for sneii and the yields of iwamoto et al .
( 1999 ) for solar - metallicity sneia ( their model w7 ) at all metallicities .
2 . nuc 2 : as above , but we interpolate between zero - metallicity and solar - metallicity snia yields ( models w70 and w7 in iwamoto et al .
3 . nuc 3 : as above , but we modify the mn yield from sneia following cescutti et al .
( 2008 ) : @xmath9 where @xmath10 and @xmath11 are the metallicities of the snia progenitor and of the sun at birth , respectively .
nuc 4 : we use the metal - dependent yields of woosley & weaver ( 1995 ) _ with fe yields halved _ for sneii and the yields of iwamoto et al .
( 1999 ) for solar - metallicity sneia ( their model w7 ) at all metallicities . 5 . nuc 5 : as above , but we interpolate between zero - metallicity and solar - metallicity snia yields ( models w70 and w7 in iwamoto et al .
1999 ) . 6 .
nuc 6 : as above , but with a metal - dependent yield of mn from sneia as in eq .
[ eq : mnz ] .
models without mn production from sneia were computed as well ( models labeled nuc 0 ) and are discussed in sect . 4.2
. the first and third choice are in common with cescutti et al .
( 2008 ) , who conclude that the mn yield from sneia must be metallicity - dependent .
their result derives from the simultaneous analysis of the behaviour of [ mn / fe ] versus [ fe / h ] in the galactic bulge , in the solar neighbourhood and in the sgr dsph . here
, we add the study of mn evolution in @xmath0cen to their survey .
notice that cescutti et al .
( 2008 ) adopted model w7 of iwamoto et al .
( 1999 ) at all metallicities .
this was done because the physics and nucleosynthesis output especially for those elements produced in the inner part of the wd of metal - free sneia are likely to be not significantly different from those at @xmath12 0 . also ,
only a tiny fraction of all sneia that explode in a galaxy forms from truly @xmath10 = 0 matter .
in fact , a few snii explosions suffice to rise the metal content of the ism from zero to non - zero . on the other hand , the adoption of an empirical law of the form suggested by cescutti et al .
( 2008 ) for mn was justified by the fact that it could reproduce the mn evolution in different objects and that mn is one of those elements produced in the external layers of the wd and therefore more dependent on the initial metallicity of the progenitor of the c - o wd ( thielemann , private communication ) . at this point , it is worth mentioning once again that badenes et al .
( 2008 ) also find that the mn yield from sneia declines with decreasing metallicity .
the metallicity dependence they suggest is the same as in eq .
[ eq : mnz ] .
their result , which springs from detailed snia modelling , is stable against variations in the initial conditions and explosion mechanisms ( either delayed detonation or deflagration ) of the models explored .
the dashed curves in fig . [ fig : lte ] show the behaviour of [ mn / fe ] versus [ fe / h ] predicted for the solar neighbourhood ( left panel ) and for @xmath0cen ( right panel ) with our models using the metallicity - dependent yields of woosley & weaver ( 1995 ) for sneii and model w7 of iwamoto et al .
( 1999 ) for sneia at all metallicities .
the theoretical predictions , in this and all the following figures , are normalized by the solar values of grevesse & sauval ( 1998 ) . here , the results of the models are compared to lte data . it can be immediately seen that , as soon as sneia start to sensibly contribute to the chemical enrichment of the ism @xmath131 ) in the solar neighbourhood and 0.1 gyr from the beginning of the star formation ( [ fe / h ] @xmath132 ) in @xmath0cen . ] , the theoretical curves begin to diverge from the observed [ mn / fe ] versus [ fe / h ] relation .
this can be interpreted as an indication that the mn yields from sneia are overestimated ( see also cescutti et al .
2008 ) . indeed , by introducing a metallicity dependence in the mn yield from sneia ( either by interpolating between models w70 and w7 of iwamoto et al .
1999 or by adopting a mn yield decreasing with decreasing metallicity as in cescutti et al .
2008 see eq .
[ eq : mnz ] ) , we get a better fit to the solar neighbourhood data ( left panel , dotted and solid lines , respectively ) .
however , the low [ mn / fe ] ratios observed in the most metal - rich stars of @xmath0cen can not be recovered by the model , which always predicts a [ mn / fe ] ratio increasing with time ( metallicity ) in @xmath0cen ( right panel , dotted and solid lines ) . the analysis of the evolution of several element - to - iron abundance ratios in the milky way led timmes , woosley & weaver ( 1995 ) to favour the use of _ half the nominal values _ of the fe yields given by woosley & weaver ( 1995 ) in chemical evolution models .
therefore , we have recomputed our set of chemical evolution models by reducing the woosley & weaver ( 1995 ) original fe yields by a factor of 2 .
the results are shown in fig .
[ fig : nlte ] , for either metal - independent ( dashed curves ) or metal - dependent ( dotted and solid curves ) mn yields from sneia .
the theoretical predictions are compared to non - lte abundance data .
the non - lte corrected mn abundances of metal - poor stars are higher than the corresponding lte values .
therefore , a very good agreement is obtained between the observed and predicted [ mn / fe ] versus [ fe / h ] trends for the galactic halo when assuming half the nominal values of the fe yields by woosley & weaver ( 1995 ) and the non - lte corrected mn abundances for halo stars . for [ fe / h ]
@xmath141.0 , a rise to solar ratios shallower than predicted should be probably preferred , though more data are needed to characterize the run of [ mn / fe ] with [ fe / h ] at disc metallicities .
the adoption of a metallicity - dependent yield of mn from sneia produces a constant rather than increasing [ mn / fe ] ratio in @xmath0cen only when the empirical law by cescutti et al .
( 2008 ) is adopted ( solid line ) .
interpolating linearly between models w70 and w7 of iwamoto et al .
( 1999 ) , instead , still results in a [ mn / fe ] ratio increasing with [ fe / h ] ( dotted line ) .
even a constant [ mn / fe ] ratio , however , does not suffice to explain the observations .
current measurements , in fact , seem to point to an abrupt fall of [ mn / fe ] at [ fe / h ] @xmath131.2 in this cluster ( fig .
[ fig : nlte ] , right panel ) , though one must be aware that this trend is dictated by just two stars . a decreasing [ mn / fe ] ratio in @xmath0cen
may be obtained by assuming that sneia do not produce any mn ; however , in that case it is impossible to reproduce the milky way data ( cfr . models labeled nuc 0 , dot - dashed lines in fig .
[ fig : nlte ] ) .
this could imply that sneia in the milky way and in @xmath0cen have different progenitors , but the paucity of mn data for @xmath0cen prevents us from drawing any firm conclusion .
the extremely low mn abundance measured for a few stars in @xmath0cen is not the only chemical peculiarity of this cluster .
some of its stars are , in fact , enormously enriched in helium and @xmath15-process elements , while characteristic anticorrelations exist among the abundances of particular elements , similarly to what is found in other globular clusters , but at variance with the milky way field at the same metallicities . in order to explain the peculiar patterns observed in the chemical abundances of a significant fraction of galactic globular cluster stars , dercole et al .
( 2008 ) have suggested that the ejecta of first - generation asymptotic giant branch ( agb ) stars collect in a cooling flow into the cluster core , where they form a subsystem of chemically anomalous second - generation stars .
though attractive , such a scenario can not provide a justification for the low mn abundances ( [ mn / fe ] @xmath130.8 ) observed in metal - rich stars of @xmath0cen by cunha et al .
( 2010 ) . indeed
, first - generation agb stars in @xmath0cen would display a value of [ mn / fe ] close to @xmath10.4 ( or @xmath10.2 , depending on whether the snii fe yields are taken at face value or reduced by a factor of 2 ) in their ejecta , dictated by snii nucleosynthesis at low metallicities .
this value is well above that suggested by the observations of relatively metal - rich stars in @xmath0cen .
the same considerations apply to the competitive scenario proposed by decressin et al .
( 2007 ) , where the chemical peculiarities of second - generation stars are driven by the slow winds of rotating massive stars .
alternatively , one might argue that chemically peculiar stars did not originate in the cluster itself , but were accreted from the surroundings ( see fellhauer , kroupa & evans 2006 , for a possible scenario of field stars trapping by the newborn @xmath0cen ) . in this case
, they would be no longer representative of the self - enrichment history of the cluster .
however , mn abundances as low as [ mn / fe ] @xmath80.8 , as measured in @xmath0cen , have never been detected elsewhere , which makes the field star capture hypothesis highly unreliable . in fig .
[ fig : all ] , we compare the [ mn / fe ] versus [ fe / h ] trends observed in different stellar systems
the solar neighbourhood , several galactic globular clusters and a few dsphs to the [ mn / fe ] versus [ fe / h ] relation of @xmath0cen stars ( notice that part of the large scatter in the @xmath0cen data has been artificially introduced by plotting together measurements based on different mn lines ; see discussion in sect . 2 and pancino et al .
it is clearly seen that , below [ fe / h ] = @xmath10.8 , all stars , independently of the system they belong to , share a common value of [ mn / fe ] of roughly @xmath10.4 .
this is not true only for a few stars in @xmath0cen with [ mn / fe ] @xmath130.8 . at this point
, it remains to be assessed whether we can explain the observations by relaxing some of the simplifying hypotheses of our model . the solid curve in fig . [ fig : disp ] , left - hand panel , tracks down the predicted behaviour of [ mn / fe ] as a function of [ fe / h ] in the ism of @xmath0cen , according to our fiducial model for this cluster . in the framework of homogeneous chemical evolution ,
the chemical composition of the newborn stars is exactly that predicted for the ism at the time of their formation .
any spread in the data is , therefore , left unexplained . for [ fe / h ] @xmath71.4 ,
the majority of the data points is consistent with the predictions of our homogeneous chemical evolution model , within the quoted uncertainties .
for [ fe / h ] @xmath141.4 , there are only 3 stars with mn determinations in @xmath0cen , and all of them lie below the theoretical curve . while this may be a hint for a decrease of [ mn / fe ] with [ fe / h ] in @xmath0cen , it was pointed out to us that one should also consider the possibility that the stars form from a medium that is not well - mixed and thus bears the signature of chemical enrichment from a few sne only . in fig .
[ fig : disp ] , right - hand panel , we show the composition of pure snii ejecta as a function of the initial stellar mass , for a grid of stellar models from 12 to 40 m@xmath16 ( filled circles and triangles ) .
we take the yields from woosley & weaver ( 1995 ) , for two values of the metallicity typical of @xmath0cen s stars .
if the star formation in @xmath0cen s progenitor is initiated by sne and the stars form from a mixture of ` snowplowed ' interstellar medium and individual supernova ejecta , in a scenario resembling that suggested for the galactic halo by ishimaru & wanajo ( 1999 ; see also tsujimoto et al . 1999 ,
argast et al .
2000 and cescutti 2008 , for inhomogeneous chemical evolution modeling of the halo ) , a broad distribution of [ mn / fe ] in the newborn stars is expected , because of the broad distribution of [ mn / fe ] ratios in the ejecta of individual core - collapse sne . however , the lowest [ mn / fe ] ratios observed in @xmath0cen are left unexplained .
also shown in fig .
[ fig : disp ] are the [ mn / fe ] ratios in the ejecta of sneia that explode in @xmath0cen s progenitor ( shaded area ; the mn yield from sneia is computed according to cescutti et al.s 2008 recipe ) . once again , it is seen that the stars with the lowest [ mn / fe ] ratios in @xmath0cen can not be explained as forming from pure sn ejecta .
though relaxing the hypothesis of homogeneous chemical evolution seems a promising way to obtain for some stars theoretical mn abundances lower than predicted for the gas , the model predictions can hardly be brought into agreement with the observations of the most metal - rich stars .
we also want to emphasize the following : inhomogeneous chemical evolution models for the galactic halo always predict that the spread in the data is reduced in the course of the evolution of the system ( see the references above ) , eventually leading to convergence with the predictions from homogeneous models .
however , in @xmath0cen we would see exactly the opposite , i.e. an increase in the dispersion with time , _ unless the data for the most metal - rich stars are tracing a genuine decrease _ of the [ mn / fe ] ratio in the cluster .
this is an intriguing aspect , that should be further investigated in more populous samples of high - metallicity @xmath0cen stars .
in this paper we present the theoretical [ mn / fe ] versus [ fe / h ] relationships predicted with our self - consistent chemical evolution model for @xmath0cen , by using different prescriptions on mn and fe synthesis in stars .
in particular , as for mn production from sneia , we adopt either a metal - independent yield , by assuming iwamoto et al.s ( 1999 ) w7 model at all metallicities , or a metal - dependent one , by interpolating between models w70 and w7 of iwamoto et al .
( 1999 ) or by adopting the empirical law of cescutti et al .
we also run models without mn production from sneia .
the theoretical relations are then compared to lte and non - lte data on mn abundances in @xmath0cen giants .
the adopted chemical evolution model reproduces all the major chemical properties of @xmath0cen its metallicity distribution function , age - metallicity relation , average trends of several @xmath17-element - to - iron abundance ratios as functions of [ fe / h ] ( romano et al .
it also accounts for the presence of extreme he - rich stars ( in the right percentage ) and for the existence of a na - o anticorrelation in the cluster ( romano et al .
however , the trend of decreasing [ mn / fe ] with increasing [ fe / h ] , displayed by both lte and non - lte data , is not reproduced by the model , which predicts instead a [ mn / fe ] ratio either increasing or constant in time , depending on the choice of stellar yields .
the adoption of a metallicity - dependent , rather than metal - independent , mn yield from sneia leads to an almost flat [ mn / fe ] versus [ fe / h ] trend in @xmath0cen only if the empirical law of cescutti et al .
( 2008 ) is adopted . by interpolating between models w70 and w7 of iwamoto et al .
( 1999 ) , i.e. between models computed for metal - free and solar - metallicity sneia , respectively , we still produce a [ mn / fe ] ratio increasing with [ fe / h ] in @xmath0cen . on the basis of the discussion in sect . 4 of the present paper and on previous results by cescutti et al .
( 2008 ) , we conclude that a flat trend has to be preferred over an increasing one . in our chemical evolution model for @xmath0cen ,
the chemical properties of the cluster are mainly driven by the action of strong differential galactic winds , which deeply affect the evolution of its dsph precursor ( see romano et al .
2007 , and references therein , for details on the adopted scenario ) .
allowing for cooling flows or field star capture would not help to explain the presence of low - mn stars in the cluster .
relaxing the hypothesis of homogeneous chemical evolution , i.e. allowing the stars to form from a mixture of ism and individual sn ejecta , could eventually lead ( for some stars ) to the prediction of mn abundances lower than predicted for the gas .
however , this would hardly accommodate the mn abundances of the two most metal - rich stars in cunha et al.s ( 2010 ) sample in a scenario of flat rather than decreasing
mn evolution with time .
we suggest that more measurements of mn in @xmath0cen stars at high metallicity are needed to finally set the issue of mn evolution in @xmath0cen .
interestingly , [ mn / fe ] in sagittarius stays almost flat ( sbordone et al . 2007 ; see fig . [
fig : all ] ) and sagittarius is likely to be the local counterpart of the accretion episode which led to the formation of @xmath0cen some 10 gyr ago ( bellazzini et al .
2008 ; georgiev et al . 2009 ; carretta et al . 2010 ) .
hence , it would be not surprising if future measurements of mn in a much bigger sample of @xmath0cen stars should reveal a constant run of mn with [ fe / h ] , rather than the decreasing trend suggested by cunha et al .
( 2010 ) on the basis of extant data . on the other hand , would future measurements confirm a fall of the [ mn / fe ] ratio towards higher metallicities , a revision of current scenarios of the formation of the cluster may be needed , since none of them is able to explain a decreasing trend of mn in @xmath0cen .
we have shown that , if mn production from sneia is totally suppressed in @xmath0cen , a [ mn / fe ] ratio decreasing with time ( metallicity ) can be found in the cluster .
while it is pointless to speculate on this theoretical result until more mn data for @xmath0cen stars become available , we note that jonhson & pilachowski ( 2010 ) observe consistently elevated [ @xmath17/fe ] ratios for nearly all stars in the cluster ( their sample totals 855 giants ) and interpret this as evidence against a significant contribution to @xmath0cen s chemical enrichment from sneia .
we thank e. carretta for providing unpublished data on mn abundances in galactic globular clusters .
one of us ( fm ) thanks prof .
thielemann for illuminating discussions on nucleosynthesis in sneia .
dr thanks v. hill and e. tolstoy for comments on preliminary results from this work presented at the international space science institute ( issi ) in bern ( ch ) in the context of team meetings .
the hospitality and financial support of issi are also gratefully acknowledged .
we acknowledge support from italian miur ( ministero dellistruzione , delluniversit e della ricerca ) through grant prin 2007 , prot .
2007jjc53x_001 and inaf through grant prin 2009 `` formation and early evolution of massive star clusters '' . 90 argast d. , samland m. , gerhard o. e. , thielemann f .- k . , 2000 , a&a , 356 , 873 arnett w. d. , 1971 , apj , 166 , 153 badenes c. , bravo e. , hughes j. p. , 2008
, apj , 680 , l33 bai g. s. , zhao g. , chen y. q. , shi j. r. , klochkova v. g. , panchuk v. e. , qiu h. m. , zhang h. w. , 2004 , a&a , 425 , 671 bekki k. , freeman k. c. , 2003 , mnras , 346 , l11 bellazzini m. , ibata r. a. , chapman s. c. , mackey a. d. , monaco l. , irwin m. j. , martin n. f. , lewis g. f. , dalessandro e. , 2008 , aj , 136 , 1147 bergemann m. , gehren t. , 2008 , a&a , 492 , 823 carretta e. , bragaglia a. , gratton r. g. , lucatello s. , bellazzini m. , catanzaro g. , leone f. , momany y. , piotto g. , dorazi v. , 2010 , apj , 714 , l7 carretta e. , bragaglia a. , gratton r. g. , lucatello s. , momany y. , 2007 , a&a , 464 , 927 carretta e. , gratton r. , cohen j. g. , beers t. c. , christlieb n. , 2002 , aj , 124 , 481 cayrel r. , depagne e. , spite m. , et al .
, 2004 , a&a , 416 , 1117 cescutti g. , 2008 , a&a , 481 , 691 cescutti g. , matteucci f. , lanfranchi g. a. , mcwilliam a. , 2008 , a&a , 491 , 401 chiappini c. , matteucci f. , gratton r. , 1997 , apj , 477 , 765 cohen j. g. , 1981 , apj , 247 , 869 cunha k. , smith v. v. , bergemann m. , suntzeff n. b. , lambert d. l. , 2010 , apj , 717 , 333 decressin t. , meynet g. , charbonnel c. , prantzos n. , ekstrm s. , 2007 , a&a , 464 , 1029 dercole a. , vesperini e. , dantona f. , mcmillan s. l. w. , recchi s. , 2008 , mnras , 391 , 825 fellhauer m. , kroupa p. , evans n. w. , 2006 , mnras , 372 , 338 feltzing s. , fohlman m. , bensby t. , 2007 , a&a , 467 , 665 freeman k. c. , 1993 , in smith g. h. , brodie j. p. , eds , asp conf .
48 , the globular clusters - galaxy connection .
, san francisco , p. 608 georgiev i. y. , hilker m. , puzia t. h. , goudfrooij p. , baumgardt h. , 2009 , mnras , 396 , 1075 gratton r. g. , 1982 , a&a , 115 , 336 gratton r. g. , carretta e. , claudi r. , lucatello s. , barbieri m. , 2003 , a&a , 404 , 187 gratton r. g. , lucatello s. , bragaglia a. , carretta e. , momany y. , pancino e. , valenti e. , 2006 , a&a , 455 , 271 greggio l. , renzini a. , 1983 , mem .
ital . , 54 , 311 grevesse n. , sauval a. j. , 1998 , space sci .
, 85 , 161 ishimaru y. , wanajo s. , 1999 , apj , 511 , l33 iwamoto k. , brachwitz f. , nomoto k. , kishimoto n. , umeda h. , hix w. r. , thielemann f .- k . , 1999 ,
apjs , 125 , 439 johnson c. i. , pilachowski c. a. , 2010 , apj , 722 , 1373 kroupa p. , tout c. a. , gilmore g. , 1993 , mnras , 262 , 545 lai d. k. , bolte m. , johnson j. a. , lucatello s. , heger a. , woosley s. e. , 2008 , apj , 681 , 1524 matteucci f. , greggio l. , 1986 , a&a , 154 , 279 matteucci f. , recchi s. , 2001 , apj , 558 , 351 mcwilliam a. , rich r. m. , smecker - hane t. a. , 2003 , apj , 592 , l21 nissen p. e. , chen y. q. , schuster w. j. , zhao g. , 2000 , a&a , 353 , 722 norris j. e. , ryan s. g. , beers t. c. , 2001 , apj , 561 , 1034 ohkubo t. , umeda h. , nomoto k. , yoshida t. , 2006 , aipc , 847 , 458 pancino e. , mucciarelli a. , sbordone l. , bellazzini m. , pasquini l. , monaco l. , ferraro f. r. , 2011 , a&a , 527 , 18 prochaska j. x. , mcwilliam a. , 2000 , apj , 537 , l57 reddy b. e. , tomkin j. , lambert d. l. , allende prieto c. , 2003 , mnras , 340 , 304 reddy b. e. , lambert d. l. , allende prieto c. , 2006 , mnras , 367 , 1329 romano d. , 2010 , in charbonnel c. , tosi m. , primas f. , chiappini c. , eds , proc .
268 , light elements in the universe .
cambridge univ . press , cambridge , p. 431
romano d. , chiappini c. , matteucci f. , tosi m. , 2005 , a&a , 430 , 491 romano d. , karakas a. i. , tosi m. , matteucci f. , 2010b , a&a , 522 , 32 romano d. , matteucci f. , tosi m. , pancino e. , bellazzini m. , ferraro f. r. , limongi m. , sollima a. , 2007 , mnras , 376 , 405 romano d. , tosi m. , cignoni m. , matteucci f. , pancino e. , bellazzini m. , 2010a , mnras , 401 , 2490 sbordone l. , bonifacio p. , buonanno r. , marconi g. , monaco l. , zaggia s. , 2007 , a&a , 465 , 815 scalo j. m. , 1986 , fundam .
cosmic phys . , 11 , 1 shetrone m. , venn k. a. , tolstoy e. , primas f. , hill v. , kaufer a. , 2003 , aj , 125 , 684 sobeck j. s. , ivans i. i. , simmerer j. a. , sneden c. , hoeflich p. , fulbright j. p. , kraft r. p. , 2006
, aj , 131 , 2949 timmes f. x. , woosley s. e. , weaver t. a. , 1995 , apjs , 98 , 617 tinsley b. m. , 1979 , apj , 229 , 1046 tsujimoto t. , shigeyama t. , yoshii y. , 1999 , apj , 519 , l63 woosley s. e. , weaver t. a. , 1995 , apjs , 101 , 181 | we model the evolution of manganese relative to iron in the progenitor system of the globular cluster omega centauri by means of a self - consistent chemical evolution model .
we use stellar yields that already reproduce the measurements of [ mn / fe ] versus [ fe / h ] in galactic field disc and halo stars , in galactic bulge stars and in the sagittarius dwarf spheroidal galaxy .
we compare our model predictions to the mn abundances measured in a sample of 10 red giant members and 6 subgiant members of @xmath0cen .
the low values of [ mn / fe ] observed in a few , metal - rich stars of the sample can not be explained in the framework of our standard , homogeneous chemical evolution model .
introducing cooling flows that selectively bring to the cluster core only the ejecta from specific categories of stars does not help to heal the disagreement with the observations .
the capture of field stars does not offer a viable explanation either .
the observed spread in the data and the lowest [ mn / fe ] values could , in principle , be understood if the system experienced inhomogeneous chemical evolution .
such an eventuality is qualitatively discussed in the present paper . however , more measurements of mn in @xmath0cen stars are needed to settle the issue of mn evolution in this cluster
. = = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = [ firstpage ] nuclear reactions , nucleosynthesis , abundances galaxies : evolution globular clusters : individual : @xmath0 centauri . |
recent progress in the processing of graphite have made possible the isolation of two dimensional carbon sheets , known as graphene @xcite .
this system has been studied theoretically for a number of years , because when rolled up they form carbon nanotubes @xcite .
the material is unique because the underlying honeycomb lattice has a band structure with dirac points at the corners of the brillouin zone , two of which are inequivalent .
undoped , the system has one electron per atom and the fermi energy surface passes directly through the dirac points . for low energies and
dopings the system may be described by the dirac equation .
the stabilization of flat graphene sheets has allowed the application of perpendicular magnetic fields and the observation of the integer quantized hall effect @xcite .
the striking result in these experiments is the first step height in the hall conductance as a function filling factor , corresponding to filling the lowest landau level ( with electrons or holes ) , is _ half _ that of all subsequent steps .
this behavior was expected @xcite based on the bulk energy spectrum of graphene in a magnetic field : for a given spin species , there are pairs of landau level bands at positive energies , each with partners at negative energies due to particle - hole symmetry . the lowest landau level ( lll )
has two levels precisely at zero energy , each of which is its own particle - hole conjugate @xcite .
this property of the lll results in its smaller contribution to the hall conductance .
the recent experiments on graphene studied ribbons that were relatively narrow , with widths in the micron @xcite or submicron @xcite range . under such circumstances , transport in the quantum hall regime
is typically dominated by edge states @xcite . in this work we study edge states for graphene ribbons in detail , focusing on the simplest cases of a zigzag edge and an armchair edge .
we demonstrate that a continuum description of edge states based on the dirac equation is possible with the adoption of appropriate boundary conditions . in an edge state description ,
the quantization of the hall conductance is determined by the number of edge state bands crossing the fermi level .
the hall conductance results imply that the lll supports only a single particle - like and a single hole - like band ( per spin ) at each edge , while the higher landau levels have twice as many .
our goal is to understand how and why this happens , in a non - interacting picture .
( because real spin plays no role in this study , we will from here on assume that all the electrons are spin - polarized , and refrain from explicitly noting the spin degree of freedom in our discussions . )
we now summarize our results . for zigzag edges , we show the correct boundary condition is for the wavefunction to vanish on a single sublattice across the edge . in this case
the lll supports two types of edge states , which we call current - carrying and dispersionless surface states .
both states have strong components at the boundary of the system , but the former has equal weights on both types of sites of the honeycomb lattice , whereas the latter exists essentially only on one sublattice and has precisely zero energy .
such zero energy surface states are well - known to exist in graphene ribbons in the absence of a magnetic field @xcite , and have been shown in tight - binding calculations to persist when a field is applied @xcite . in the quantum hall context
we find that the dispersionless surface states play a special role in forming two branches of edge states that do not pass through the fermi level for any non - zero doping .
the current - carrying edge states are also remarkable in that the surface contribution on one sublattice can be highly separated from a component well - inside the bulk of the sample , on the other sublattice
. an interesting consequence of this is that a particle injected near a zigzag edge should oscillate back and forth between the edge and bulk , although presumably such oscillations would be damped by many - body effects not included in our study . for armchair edges
, we find the correct boundary condition is vanishing of the wavefunction on both sublattices at the edge .
this is achieved by mixing of wavefunctions from both dirac points .
in this case , there are no dispersionless surface states , and the lll edge states behave differently than the higher landau levels for other reasons .
as we shall see , the energetics of states from one of these valleys is generically higher than from the other on a given sublattice , so that in the lll only one band of edge states can meet the boundary condition , whereas in higher landau levels there are two such bands .
the admixing of the two valleys leads to wavefunctions with a characteristic spatial oscillation of the electron density with wavevector equal to @xmath0 , the difference between the dirac point wavevectors along the direction perpendicular to the edge .
such oscillations should be observable in stm measurements .
we begin by reviewing some generalities about electrons in graphene .
the lattice structure is a triangular lattice whose primitive lattice vectors are @xmath1 and @xmath2 .
there are two atoms per unit cell located at @xmath3 and at @xmath4 .
a simple tight - binding model with only nearest neighbor hopping @xmath5 leads to a hamiltonian with dirac points at the six corners of the brillouin zone , only two of which are inequivalent , and we take these to be @xmath6 and @xmath7 .
wavefunctions can be expressed via the @xmath8 approximation @xcite in terms of envelope functions @xmath9 $ ] and @xmath10 $ ] for states near the @xmath11 and @xmath12 points , respectively .
these can be conveniently combined into a 4-vector @xmath13 .
( the reason for this sign convention will become apparent when we discuss the armchair edge . ) this satisfies a dirac equation @xmath14 , @xmath15 with @xmath16 . note that @xmath17 denotes the separation in reciprocal space of the wavefunction from the @xmath11 ( @xmath12 ) point in the upper left ( lower right ) block of the hamiltonian .
to apply this hamiltonian in the presence of a magnetic field , one makes @xcite the peierls substitution @xmath18 where @xmath19 is the vector potential . before applying this procedure to systems with an edge
, we point out some interesting and useful properties of @xmath20 .
firstly , @xmath20 ( and the more exact tight - binding hamiltonian from which it descended ) has @xcite chiral ( i.e. , particle - hole ) symmetry , @xmath21 , where @xmath22 , and @xmath23 is the pauli matrix .
this tells us that a solution to the dirac equation @xmath24 with energy @xmath25 has a particle - hole conjugate partner @xmath26 with energy @xmath27 .
because of this , the wavefunctions must be normalized on each sublattice separately : @xmath28= 1/2 $ ] , for @xmath29 .
the solutions for states well away from the edge are well - known @xcite . taking @xmath30 ,
@xmath31 , and @xmath32 with @xmath33 the @xmath34 extension of the sample , the wavefunctions retain their valley index as a good quantum number , and the positive energy wavefunctions may be written as @xmath35 $ ] for the @xmath11 valley , and @xmath36 $ ] for the @xmath12 valley , with energies @xmath37 . in these expressions
, @xmath38 is the @xmath39th harmonic oscillator state and @xmath40 is the magnetic length .
the negative energy states are easily obtained by reversing the signs of the wavefunctions on the b sublattice . for the case of the lll , @xmath41 , and
only one component of the 4-vectors for each valley is non - zero .
this means the particle - hole conjugate of these wavefunctions are themselves .
the bulk lll wavefunctions do not have a clear particle- or hole - like character .
the geometry for a zigzag edge is illustrated on the top and bottom edges of fig .
1 . it is interesting to note that each atom at the edge is of the same sublattice ( say a ) .
we shall see below that the appropriate boundary condition is to set the wavefunction to zero on a single sublattice ( b ) , which we can understand to be the line of lattice sites that would lie just above or below the system if the bonds had not been cut to form the edge . in our discussions
we will work with edges that lie along the @xmath34 direction , so in what follows the coordinate axes in fig .
1 will be rotated by 90@xmath42 .
we begin by computing the band structure for a tight binding model of a graphene ribbon with zigzag edges , an example of which is illustrated in fig .
the flat degenerate bands over a range of @xmath43 are landau levels , and , in the case of the lll , dispersionless surface states which we discuss below . in a wide sample
, there is generically a large degeneracy within each landau band , because for the @xmath44 valley there are wavefunctions peaked at @xmath45\ell^2 $ ] , where @xmath46 is a reciprocal lattice vector for the ribbon , and the integer @xmath39 can take any value such that @xmath47 is between the sample edges .
one may conveniently reorganize the states by allowing all values of @xmath43 such that @xmath48\ell^2 < l_x/2 $ ] , with @xmath49 the ribbon width , and assigning one state for each @xmath43 in the extended zone . the prominent structure in fig .
2 is the appearance of dispersing energy bands , which occur when the wavefunctions approach the edges . for the higher landau levels
one observes two pairs of such bands , whereas for the lll there is only one such pair .
this means that a fermi energy crossing between the @xmath39th and @xmath50th landau levels yields a hall conductance @xmath51 , as observed in experiment @xcite .
the unique behavior of the lll edge states may be understood in terms of eigenstates of the dirac hamiltonian with vanishing boundary conditions on a single sublattice .
we begin by rotating the wavevectors in eq .
[ hamilt_kp ] , @xmath52 so that the zigzag edge lies along the @xmath34 , and our wavefunctions then exist in the space @xmath53 . taking @xmath30 in this coordinate system , and defining the ladder operator @xmath54 $ ] , with @xmath55 for the @xmath44 valley , the wavefunctions obey @xmath56
@xmath57 it is easy to see if one solves the equations for @xmath58 and @xmath59 , the remaining wavefunctions are determined by @xmath60 and @xmath61 .
for the zigzag edge , the boundary condition does not admix valleys , and we can meet it for each type of wavefunction separately : @xmath62 .
thus for the * k * valley the spectrum @xmath63 is identical to that of a quantum hall edge with a sharp boundary @xcite .
the wavefunctions @xmath64 similarly are identical to their standard hall edge counterparts , turning into states in the @xmath39th landau level as @xmath65 moves well away from the edge . for @xmath66
, @xmath67 is quite similar to a state in the @xmath68 landau level provided the center of the wavefunction is not too close to the edge . in the lll ( @xmath41 ) , @xmath69 is qualitatively different .
because lll states are annihilated by the ladder operator @xmath70 , and @xmath58 is similar to a bulk lll state when the center of the wavefunction is not too close to the edge , @xmath69 is extremely small , except close to @xmath71 where @xmath58 vanishes and is forced to deviate from a bulk lll state .
the result is that @xmath69 is strongly confined to the surface , and as the center of the @xmath58 moves further into the interior of the sample , @xmath69 becomes increasingly so confined . despite this strongly localized form
, the normalization of the wavefunctions discussed in the previous section requires that fully half the probability of finding the electron resides in this surface contribution .
we note that , because @xmath25 disperses with @xmath43 , these surface states carry current and contribute to the hall conductivity . within the dirac equation , the existence and form of these current - carrying surface states can be examined with a variational approach .
we adopt a trial wavefunction @xmath72 and require @xmath73 .
one may easily confirm that @xmath74 , where @xmath75 .
a simple choice for @xmath76 is @xmath77 , where @xmath78 is a normalization constant and our variational parameter is @xmath79 . with
this choice one finds @xmath80 ^ 2/2\ell^2}\rbrace$ ] , so that if @xmath81 , @xmath69 is confined to the surface .
the result of minimizing @xmath82 is illustrated in fig .
3 , where in the inset one sees that @xmath79 increases faster than @xmath43 , so that @xmath69 becomes more confined to the surface of the sample as @xmath43 increases and @xmath58 penetrates into the bulk .
direct examination of wavefunctions from the tight - binding model confirms this basic picture , except in one important respect . whereas the dirac equation allows current carrying states with @xmath69 increasingly localized to the surface as @xmath43 grows ( and the peak position @xmath47 of @xmath58 moves into the bulk ) ,
the tight - binding results show that for @xmath83 the lll wavefunction on sublattice a and the surface state on sublattice b appear as separate states , and that the surface state now moves back into the interior of the system with further increase in @xmath43 .
the current - carrying surface state evolves into a dispersionless surface state which we describe below , and the number of allowed current - carrying surface states is limited .
this is clearly an effect of the discreteness of the lattice that is not captured by the dirac equation : independent , highly localized surface states can be written down for any @xmath43 in the continuum , but on the lattice states at @xmath43 and @xmath84 have the same periodicity along the @xmath34 direction , and so can not be made independent .
thus the number of current - carrying surface states is limited to @xmath85 in the allowed @xmath43 s for the current - carrying surface states . for @xmath86
, one may take the @xmath11 valley wavefunction to have its bulk form .
the edge state wavefunctions of the @xmath12 valley on the a sublattice are analogously identical to the well - known ones @xcite , although @xmath87 is shifted upward by a single unit ( @xmath88 ) due to the ordering of the operators in the last of eqs .
[ shop ] . for the lll ,
the bulk states @xmath89 , @xmath90 , with @xmath91 , exactly satisfy eqs .
[ shop ] with zero energy .
remarkably , these states are unaffected by the edge . moreover , because @xmath92 , there are values values of @xmath43 in the extended zone for which @xmath93 , and the state is confined to the surface .
these are the dispersionless surface states : they do not contribute to the hall conductivity .
the lll wavefunctions of the tight - binding results around the center of the bands in fig .
2(a ) behave precisely as the dirac equation results suggest : they are strongly confined to the surface , and continuously evolve into bulk lll states on the a sublattice as @xmath43 increases .
the dispersionless surface states supported by the lll are the reason that it carries only half the hall conductivity of the higher landau levels for the zigzag edge @xcite .
the armchair edge is illustrated as the left and right edges fig . 1 , and the corresponding bandstructure from a representative tight - binding calculation appears in fig 2(b ) .
here the edge runs along the @xmath34 direction , and no rotation of the figure is needed to represent our calculations . unlike the zigzag edge , the landau bands all have dispersing states in the same regions of @xmath43 , but the lll has one pair each of hole - like and particle - like edge states , while all the higher landau levels have two . to understand this from the viewpoint of the dirac equation , we need to impose appropriate boundary conditions . in fig . 1
one may see that the termination consists of a line of a - b dimers , so it is natural to have the wavefunction amplitude vanish on both sublattices at @xmath71 . to do this
we must admix valleys , and require @xmath94 and @xmath95 . using the dirac equation , and the fact that @xmath96 , this second condition implies @xmath97 . to understand the effect of this on the solutions , it is convenient to combine the @xmath58 s into a single wavefunction defined for @xmath98 : @xmath99 , with @xmath100 the step function .
the boundary conditions then amount to @xmath101 and its derivative being continuous at @xmath71 .
( this was the reason for our choice of relative sign in the 4-vectors of section ii . ) from eqs .
[ sho ] and [ shop ] , it is easy to see that @xmath102 obeys a schroedinger equation @xmath103\psi(x)=\tilde{\varepsilon}^2\psi(x)$ ] , with @xmath104 $ ] . for large @xmath65 ,
this double well potential , illustrated in fig .
4 , has low energy states associated with the left well at @xmath105 , while for the right well one has states at @xmath106 .
we thus see there will be hybridization leading to pairs of edge states for all the higher landau levels , whereas for the lll there will be just a single such state .
the admixing of different valley states to meet the boundary condition means that the wavefunction will oscillate with period @xmath107 .
the behavior can explicitly be seen in fig . 5 , which illustrates a lll edge state from the tight - binding calculation .
the apparent oscillation has precisely the period one expects for the valley mixing we introduced in the dirac equation to meet the boundary condition .
although the period of this oscillation is very short ( 3.69 ) , it is in principle observable by stm measurements because the samples can be open to their environment @xcite , in contrast to gaas systems .
in this paper we have studied the edge states of graphene ribbons with zigzag and armchair terminations .
we found in both these cases that a continuum description in the form of the dirac equation captures most features of the states found in tight - binding calculations , provided the wavefunction vanishes at the termination of the sample .
for zigzag edges , we found the boundary condition can be met by wavefunctions within a single valley , leading to two types of edge states in the lowest landau level , current - carrying surface states and dispersionless surface states .
the latter of these explains why the contribution to the quantized hall coefficient from the lll is only half that of higher landau levels . for the armchair edge
, we found that admixing of valleys is necessary to satisfy the boundary condition . for higher landau levels , there are two hybridizations of the valley states for which this is possible , whereas in the lll there is only one .
this again leads to a contribution to the hall coefficient from the lll half the size of those from other occupied landau levels .
terms that break the particle - hole symmetry ( e.g. , next nearest neighbor hopping ) can introduce an energy dependence in the dispersionless surface states
. however , the resulting edge states are not chiral and do not contribute to the hall conductivity .
see ref . . | we study edges states of graphene ribbons in the quantized hall regime , and show that they can be described within a continuum model ( the dirac equation ) when appropriate boundary conditions are adopted .
the two simplest terminations , zigzag and armchair edges , are studied in detail . for zigzag edges , we find that the lowest landau level states terminate in two types of edge states , dispersionless and current - carrying surface states .
the latter involve components on different sublattices that may be separated by distances far greater than the magnetic length . for armchair edges ,
the boundary conditions are met by admixing states from different valleys , and we show that this leads to a single set of edges states for the lowest landau level and two sets for all higher landau levels . in both cases ,
the resulting hall conductance step for the lowest landau level is half that between higher landau levels , as observed in experiment . |
left - right symmetric models ( lrsm ) @xcite provide a framework within which spontaneous parity breaking as well as tiny neutrino masses @xcite can be successfully implemented without reference to very high scale physics such as grand unification .
incorporating supersymmetry ( susy ) into it comes with other advantages like providing a solution to the gauge hierarchy problem , and providing a cold dark matter candidate which is the lightest supersymmetric particle ( lsp ) . in minimal supersymmetric standard model ( mssm ) ,
the stability of lsp is guaranteed by r - parity , defined as @xmath4 where @xmath5 is the spin of the particle .
this is a discrete symmetry put by hand in mssm to keep the baryon number ( b ) and lepton number ( l ) violating terms away from the superpotential . in generic implementations of left - right symmetry ,
r - parity is a part of the gauge symmetry and hence not ad - hoc like in the mssm . in one class of models @xcite
, spontaneous parity breaking is achieved without breaking r - parity .
this was not possible in minimal supersymmetric left right(susylr ) models where the only way to break parity is to consider spontaneous r - parity violation @xcite .
in minimal susylr model parity , @xmath6 gauge symmetry as well as r - parity break simultaneously by the vacuum expectation value of right handed sneutrino .
spontaneous breaking of exact discrete symmetries like parity ( which we shall denote as d - parity hereafter ) , as well as r - parity have got cosmological implications since they lead to frustrated phase transitions leaving behind a network of domain walls ( dw ) .
these domain walls , if not removed will be in conflict with the observed universe @xcite .
it was pointed out @xcite that planck scale suppressed non - renormalizable operators can be a source of domain wall instability .
supersymmetry dictates the structure of these non - renormalizable terms but also gives rise to the gravitino overabundance problem . incorporating all these restrictions , the constraint on the d - parity breaking scale in r - parity conserving
susylr models @xcite has been discussed in @xcite . here
we extend the analysis to a more general class of models where both r - parity and d - parity break spontaneously .
it should be mentioned that the formation of domain walls is not generic in all left - right models .
models where d - parity and @xmath6 gauge symmetry are broken at two different stages do not suffer from this problem @xcite . in these models ,
the vacuum expectation value ( vev ) of a parity odd singlet field breaks the d - parity first and @xmath6 gauge symmetry gets broken at a later stage by either higgs triplets and higgs doublets .
this paper is organized as follows . in section [ sec : dw - section ] we briefly review the domain wall dynamics . in section [ sec : susylr1 ]
we discuss minimal susylr model with higgs triplets , constraints on the symmetry breaking scale from successful removal of domain wall as well as gauge coupling unification and in section [ sec : susylr2 ] we do this same analysis for minimal susylr model with higgs doublets .
we summarize our results in section [ sec : con ] .
discrete symmetries and their spontaneous breaking are both common instances and desirable in model building .
the spontaneous breaking of such discrete symmetries gives rise to a network of domain walls leaving the accompanying phase transition frustrated @xcite .
the danger of a frustrated phase transition can therefore be evaded if a small explicit breaking of discrete symmetry can be introduced . due to the smallness of such discrete symmetry breaking
, the resulting domain walls may be relatively long lived and can dominate the universe for a long time . since this will be in conflict with the observed universe , these domain walls need to disappear at a very high energy scale ( at least before big bang nucleosynthesis ) . keeping this in mind , we summarize the three cases of domain wall dynamics discussed in @xcite , one of which originates in radiation dominated ( rd ) universe and destabilized also within the radiation dominated universe .
this scenario was originally proposed by kibble @xcite and vilenkin @xcite .
the second scenario was essentially proposed in @xcite , which consists of the walls originating in a radiation dominated phase , subsequent to which the universe enters a matter dominated ( md ) phase , either due to substantial production of heavy unwanted relics such as moduli , or simply due to a coherent oscillating scalar field .
the third one is a variant of the md model in which the domain walls dominate the universe for a considerable epoch giving rise to a mild inflationary behavior or weak inflation ( wi ) @xcite . in all these cases
the domain walls disappear before they come to dominate the energy density of the universe .
when a scalar field @xmath7 acquires a vev at a scale @xmath8 at some critical temperature @xmath9 , a phase transition occurs leading to the formation of domain walls .
the energy density trapped per unit area of such a wall is @xmath10 .
the dynamics of the walls are determined by two quantitites , force due to tension @xmath11 and force due to friction @xmath12 where @xmath13 is the average scale of radius of curvature prevailing in the wall complex , @xmath14 is the speed at which the domain wall is navigating through the medium and @xmath15 is the temperature .
the epoch at which these two forces balance each other sets the time scale @xmath16 .
putting all these together leads to the scaling law for the growth of the scale @xmath17 : @xmath18 the energy density of the domain walls goes as @xmath19 . in a radiation dominated era
this @xmath20 is comparable to the energy density of the universe @xmath21 $ ] around time @xmath22 .
the pressure difference arising from small asymmetry on the two sides of the wall competes with the two forces @xmath23 and @xmath24 discussed above . for @xmath25 to exceed either of these two quantities before @xmath26 @xmath27 similar analysis in the matter dominated era , originally considered in @xcite begins with the assumption that the initially formed wall complex in a phase transition is expected to rapidly relax to a few walls per horizon volume at an epoch characterized by hubble parameter value @xmath28 .
thus the initial energy density of the wall complex is @xmath29 .
this epoch onward the energy density of the universe is assumed to be dominated by heavy relics or an oscillating modulus field and in both the cases the scale factor grows as @xmath30 .
the energy density scales as @xmath31 .
if the domain wall ( dw ) complex remains frustrated , i.e. its energy density contribution @xmath32 , the hubble parameter at the epoch of equality of dw contribution with that of the rest of the matter is given by @xcite @xmath33 assuming that the domain walls start decaying as soon as they dominate the energy density of the universe , which corresponds to a temperature @xmath34 such that @xmath35 , the above equation gives @xmath36 under the assumption that the domain walls are formed at @xmath37 @xmath38 now from eq .
( [ td ] ) @xmath39 demanding @xmath40 leads to @xmath41 the third possibility is the walls dominating the energy density of the universe for a limited epoch which leads to a mild inflation .
this possibility was considered in @xcite . as discussed in @xcite ,
the evolution of energy density of such walls can be expressed as @xmath42 where @xmath43 is the scale factor at which domain walls start dominating ( decaying ) and @xmath44 is the corresponding time .
if the epoch of domain wall decay is characterized by temperature @xmath34 , then @xmath45 and the above equation gives @xmath46 in the matter dominated era the energy densit of the moduli fields scale as @xmath47 using this in equation ( [ td1 ] ) gives @xmath48 domain walls start dominating the universe after the time of equality , @xmath49 .
so the pressure difference across the walls when they start decaying is given by @xmath50 where @xmath51 . replacing the value of @xmath52 from equation ( [ heq ] ) ,
the pressure difference becomes @xmath53 unlike the previous two cases rd and md , here it will not be possible to estimate @xmath34 in terms of other mass scales and we will keep it as undetermined .
we consider the minimal susylr model of kuchimanchi et al @xcite in this section .
although the minimality of the higgs content is an attractive feature of this model , the authours concluded that d - parity can be spontaneously broken only at the expense of breaking r - parity spontaneously at the same energy scale by the vev of right handed sneutrino .
since the r - parity violation(rpv ) is in the leptonic sector only , the dangerous proton decay problem can be evaded in this model .
the models where left handed sneutrino vev gives rise to rpv are strongly disfavored by electroweak precision measurements @xcite .
however there is no such strict constraints on models where right handed sneutrino vev gives rise to rpv provided the extra gauge boson masses lie above the allowed lower bounds @xcite .
here we find another constraint on this rpv scale from domain wall removal as well as gauge coupling unification .
the matter content of this model is @xmath54 @xmath55 the higgs sector of this minimal consists of the higgs bidoublets and higgs triplets @xmath56 @xmath57 @xmath58 the renormalizable superpotential is @xmath59 where @xmath60 and @xmath61 are symmetric matrices .
it has been shown that with this minimal field content it is not possible to break the d - parity spontaneously . adding a parity odd singlet
also does not improve the situation .
the authors showed that the d - parity breaking vacua in this case also give rise to the breaking of electromagnetic charge .
the authors @xcite proposed an alternative scenario where it was shown that by allowing a non - zero vev for right handed sneutrino , @xmath62 it is possible to get d - parity breaking minima which preserve electromagnetic charge .
however , the vev of sneutrino which has odd @xmath63 charge also gives rise to spontaneous r - parity violation .
here we follow the approximations adopted by the authors to find a region in parameter space of the coupling constants giving rise to the desired minima .
the first approximation is the one where they choose the parameter space , such that @xmath64 and @xmath65 are smaller than the constants @xmath66 and @xmath67 . with this approximation
the @xmath68-terms become weaker than the trilinear terms that contain the triplet scalars and the sleptons .
the second approximation is made in order to maintain the hierarchy between the electroweak scale and the parity breaking scale , i.e. @xmath69 . with these approximations the scalar potential can now be written as , @xmath70 + 4 \lvert f \rvert^2(\lvert \tilde{l}^t_c\tau_2 \delta_c \rvert^2+\lvert \tilde{l}^t\tau_2 \delta \rvert^2)+m'^2 \text{tr } ( \delta \bar{\delta}+\delta_c \bar{\delta}_c + h.c . ) \nonumber \\ & & + [ \tilde{l}^t\tau_2 ( iv\delta+im^*f\bar{\delta}^{\dagger } ) \tilde{l}+\tilde{l}^t_c\tau_2 ( iv\delta_c+im^*f\bar{\delta}^{\dagger}_c ) \tilde{l}_c + h.c . ]
\label{eq : scalar - potential}\end{aligned}\ ] ] consider the case where the right handed fields getting a non - zero vev , at the same time the vev for the left handed fields is zero .
so , @xmath71 @xmath72 now putting these vev s in the scalar potential eq.([eq : scalar - potential ] ) gives rise to @xmath73 minimising the above potential the authors get the solution for the parity breaking minima as @xmath74 since the original theory is left - right symmetric there exists an equivalent minima corresponding to @xmath75 .
the degeneracy of these two equivalent vacua leads to the unavoidable consequences of formation of domain walls .
for a successful phase transition accompanying the symmetry breaking , these domain walls should be unstable .
we follow the idea@xcite where it is argued that planck sale suppressed non - renormalizable operators can potentially solve the domain wall problem .
we adopt the technique developed in@xcite to find the operators suppressed by planck scale .
and we find the constrains on the symmetry breaking scale from cosmological considerations .
we now find the @xmath76 terms in the effective potential by expanding the k@xmath77hler potential and superpotential in powers of @xmath76 .
we include the terms containing @xmath78 in the expansion .
the k@xmath77hler potential in this model upto @xmath76 is @xmath79+\text{tr}[\delta_c \delta^{\dagger}_c + \bar{\delta}_c \bar{\delta}^{\dagger}_c]+\frac{c_l}{m_{pl}}(l^t\tau_2 \delta l + l^t\tau_2 \bar{\delta}^{\dagger } l + h.c . )
\nonumber \\ & & + \frac{c_r}{m_{pl}}(l^t_c\tau_2 \delta_c l_c + l^t_c\tau_2 \bar{\delta}^{\dagger}_c l_c + h.c . ) \end{aligned}\ ] ] the superpotential upto the powers of @xmath76 is @xmath80)^2 + \frac{a_r}{2m_{pl } } ( \text{tr } [ \delta_c \bar{\delta}_c])^2 + \frac{b_l}{m_{pl } } \text{tr}[\delta^2]\text{tr } [ \bar{\delta}^2]+\frac{b_r}{m_{pl } } \text{tr}[\delta_c^2 ] \text{tr}[\bar{\delta}_c^2 ] } \nonumber \\ & & + \frac{f_1}{m_{pl } } ( \text{tr } [ \delta \bar{\delta } ] ) ( \text{tr } [ \delta_c \bar{\delta}_c])+\frac{f_2}{m_{pl}}\text{tr } [ \delta^2]\text{tr}[\delta_c^2]+\frac{f_3}{m_{pl}}\text{tr } [ \bar{\delta}^2]\text{tr}[\bar{\delta}_c^2 ] + \frac{f_4}{m_{pl } } ( l^t l)(l_c^tl_c ) \nonumber\end{aligned}\ ] ] assuming a phase where only right type fields get non - zero vev and left type fields get zero vev , the scalar potential upto the leading term in @xmath76 becomes , @xmath81 using eq.([eq : pbs ] ) in the above equation we get @xmath82 similarly assuming non - zero vev for left type fields only and not for right type fields the effective potential becomes , @xmath83 if the scale of parity breaking is @xmath8 then @xmath84 . in this case where we consider the equal chance for left and right type fields getting a vev , then @xmath85 .
so the effective energy difference arising from the operators is given by , @xmath86 now we shall compare this @xmath25 with the case in a matter dominated era where we have calculated the energy density for the domain wall to decay . before going
further we make some approximations .
the constants @xmath87 and @xmath88 in eq.([eq : eff - delta - rho ] ) are the coefficients appearing in the trilinear terms in eq.([eq : scalar - potential ] ) .
these trilinear terms are the soft terms .
so the coefficients are of electroweak scale , @xmath89 .
@xmath90 and @xmath91 appear in the mass terms for @xmath92 and @xmath93 . since these triplet fields get their vev s at a scale @xmath8 , the scales @xmath90 and @xmath91 should be of order @xmath8 .
but the scale of @xmath8 is higher than the electrowek scale , hence we find @xmath94 .
so @xmath95 so the dominant term in eq.([eq : eff - delta - rho ] ) is , @xmath96 now by comparing , @xmath97 putting the electroweak scale as @xmath98gev and taking @xmath99 as @xmath100 we get the constraint on @xmath101 as @xmath102 comparing the obtained @xmath25 with the case in a radiation dominated era we get , @xmath103 proceeding as above we get the constraint on @xmath101 as @xmath104 taking the dimensionless parameters @xmath105 to be of order one , the equation [ eq : m_r - constrnt - md ] gives an upper bound on the scale @xmath8 in a matter dominated era @xmath106 similarly during the radiation dominated era , the equation [ eq : m_r - constrnt - rd ] gives an upper bound on @xmath8 @xmath107 allowing further fine tuning between @xmath108 and @xmath109 will make this bound even more strict .
however as we will see below , such a low intermediate d - parity breaking scale is not favored from successful gauge coupling unification point of view which makes this model less attractive . comparing the obtained @xmath25 with the weak inflation case we have @xmath110 taking the dimensionless coefficients to be of order one
, we arrive at the following bound on @xmath8 @xmath111 thus for @xmath34 of the order of electroweak scale , @xmath8 remains just below the gravitino bound .
however , if @xmath112 , then the @xmath8 is forced to be higher that @xmath113 which , as noted in @xcite can be problematic if the reheating temperature after the domain wall disappearance is comparable to the temperature scale of original phase transition . in that case , the universe would reheat to a temperature higher than @xmath113 giving rise to gravitino overabundance .
successful gauge coupling unification at scale @xmath114 gev puts tight constraints on the intermediate symmetry breaking scales .
assuming the higgs triplets to be as heavy as the the scale @xmath8 , we get a lower bound on the scale @xmath8 to be of the order of @xmath115 gev . a lower value of @xmath8 will make the couplings of @xmath63 and @xmath116 meet before the allowed unification scale from proton decay constraints .
although for the minimal particle content , the @xmath117 couplings do not meet the other two at one point , we can always take into account of some additional fields in a grand unified theory ( gut ) like @xmath2 which survive the gut symmetry breaking and can be as light as the scale @xmath8 . here
we consider three pairs of extra heavy colored superfields @xmath118 which can be naturally fitted within @xmath2 gut theory inside the representations @xmath119 .
the resulting unification is shown in fig .
thus the lower limit on @xmath8 from unification is in conflict with the bounds from domain wall removal ( [ mdbound]),([rdbound ] ) .
tev , @xmath120 gev ]
spontaneous r - parity breaking can be achieved even without giving vev to the sneutrino fields .
if the @xmath63 symmetry is broken by a higgs field which has odd @xmath121 charge then r - parity is spontaneously broken .
we call this model as minimal higgs doublet ( mhd ) model . the minimal such model
@xcite has the following particle content @xmath122 @xmath123 @xmath124 @xmath125 where the numbers in brackets correspond to the quantum numbers corresponding to @xmath126 .
the symmetry breaking pattern is @xmath127 the renormalizable superpotential relevant for the spontaneous parity violation is given as follows @xmath128 the scalar potential is @xmath129 where @xmath130 is the f - term scalar potential , @xmath131 is the d - term of the scalar potential and @xmath132 is the soft supersymmetry breaking scalar potential .
we introduce the soft susy breaking terms to check if they alter relations between various mass scales in the model .
the soft susy breaking superpotential in this case is given by @xmath133+h.c . )
\nonumber \\ & & + ( a_1 h^{\dagger } \phi_i h_c + a_2 \bar{h}^{\dagger } \phi_i \bar{h}_c + h.c . ) \label{soft}\end{aligned}\ ] ] where all the parameters @xmath134 are of the order of susy breaking scale @xmath135 .
we denote the vev of the neutral components of @xmath136 as @xmath137 minimizing the potential with respect to @xmath138 , we get the relations @xmath139 @xmath140 from the above two equations we arrive at @xmath141 assuming @xmath142 the above relation gives the parity breaking solution ( @xmath143 ) @xmath144 from the above relations we can show that parity is broken spontaneously to give rise to the parity violating standard model .
also there is a seesaw between @xmath145 and @xmath146 from the above equation which can give rise to tiny neutrino masses @xcite .
tev , @xmath147 gev , @xmath148 gev ] similar to the previous section , here also we find the @xmath76 terms in the effective potential by expanding the k@xmath77hler potential and superpotential in powers of @xmath76 .
the superpotential upto the powers of @xmath76 is @xmath149 the k@xmath77hler potential in this model upto @xmath76 is @xmath150 assuming a phase where only right type fields get non - zero vev and left type fields get zero vev , the scalar potential upto the leading term in @xmath76 becomes @xmath151 similarly for the phase where only left type fields get non - zero vev @xmath152 taking @xmath153 the effective energy difference can now be calculated as @xmath154\frac{m_r^5}{m_{pl}}\ ] ] thus for the matter dominated era we have @xmath155 > \frac{m^{1/2}_r}{m^{1/2}_{pl}}\ ] ] and for the radiation dominated era @xmath155 > \frac{m_r}{m_{pl}}\ ] ] taking the various dimensionless parameters to be of order one , we get the same upper bound on @xmath8 in both the above cases @xmath156 similarly , in the weak inflation scenario we have @xmath155\frac{m_r^5}{m_{pl } } \geq m^4_r \frac{t^{12}_d m^3_{pl}}{m^{15}_r}\ ] ] assuming the dimensionless coefficients to be order one , this leads to @xmath157 here @xmath34 can be as high as @xmath158 for @xmath8 to remain below the gravitino bound . as noted in the previous section , @xmath159 can lead to gravitino overabundance if the reheat temperature after the wall disappearance is same as the temperature of the original phase transition .
thus the scale @xmath8 is less restrictive in this model compared to the susylr model with higgs triplets . as we will see below
, one can have successful gauge coupling unification in this model for @xmath160 gev . due to the possibility of successful removal of domain walls as well as successful gauge coupling unification , this model is the preferred one over the model with higgs triplets .
similar to the minimal susylr model with higgs triplets , here also the intermediate symmetry breaking scales will be constrained by demanding successful gauge coupling unification at a very high scale @xmath161 .
similar to the previous case , here also the couplings of @xmath63 and @xmath116 meet much before the allowed unification scale if the intermediate symmetry breaking scale @xmath8 is lower than a certain value . for the minimal susylr model with higgs doublets ,
this lower bound on @xmath8 is found to be of the order of @xmath162 gev .
we also consider two additional heavy colored superfields so that the @xmath117 coupling meet the other two couplings at one point .
they are denoted as @xmath163 and can be accommodated within @xmath2 gut theory in the representations @xmath164 .
here we assume that the structure of the gut theory is such that these fields survive the symmetry breaking and can be as light as the @xmath6 breaking scale .
the resulting gauge coupling unification as shown in the figure [ fig2 ] .
we have discussed the issue of domain wall formation due to the spontaneous breaking of d - parity in two different versions of supersymmetric left right models : one with higgs triplets having @xmath121 charge @xmath165 and the other with higgs doublets having @xmath121 charge @xmath166 . since stable domain walls are in conflict with cosmology , we consider the effects of planck scale suppressed operators in destabilizing them .
we consider the evolution and decay of domain walls in two different epochs : radiation dominated as well as matter dominated .
we find that successful removal of domain walls put rather strict constraints on the d - parity breaking scale @xmath8 in the model with higgs triplets .
the model with higgs doublets is far less restrictive on the other hand .
.bounds on @xmath167 in r - parity violating susylr models [ cols="^,^,^,^,^ " , ] [ table2 ] we also find the constraint on the d - parity breaking scale in these models by demanding gauge coupling unification at a scale @xmath168 gev .
we use one - loop beta functions for both the models and take into account of some heavy colored superfields to make the @xmath117 coupling meet the other two exactly at one point .
the results are shown in table ( [ table1 ] ) .
we also mention the model by bhupal dev and mohapatra ( bdm ) @xcite in the table where the scale of @xmath169 symmetry breaking to @xmath170 denoted by @xmath8 can be as low as few tev .
however there is no constraint on the scale @xmath8 from domain wall disappearance due to the existence of a parity odd singlet which , after acquiring a vev breaks the degeneracy between two possible vacua . in table
( [ table2 ] ) we summarize the results of similar analysis obtained for r - parity conserving susylr models in some of our earlier works .
the bounds from domain wall disappearance in aulakh - bajc - melfo - rasin - senjanovic ( abmrs ) @xcite model and babu - mohapatra ( bm ) @xcite model were discussed in @xcite .
the bitriplet model @xcite does not suffer from domain wall problem due to the existence of a parity odd singlet as pointed out in the introduction .
the bounds on @xmath8 from gauge coupling unification in such models were discussed in @xcite . to summarize the result of this paper , it is shown that both domain wall removal and unification constraints can be satisfied in the minimal higgs doublet model or the mhd model whereas it is not possible to have successful removal of domain walls and gauge coupling unification together in the model with higgs triplets ( mslrm ) .
we would like to thank prof urjit a. yajnik , iit bombay for useful comments and discussions . | domain wall formation is quite generic in spontaneous left - right parity ( d - parity ) breaking models .
since they are in conflict with cosmology , we need some mechanisms to remove them .
planck scale suppressed effects have been considered to be quite successful for this purpose .
we study this possibility in minimal supersymmetric left - right ( susylr ) model originally proposed by kuchimanchi et al @xcite where both d - parity and r - parity @xmath0 are spontaneously broken .
we find that planck scale suppressed terms allowed for the specific particle content of this model can successfully remove the domain walls provided the d - parity breaking scale is relatively low @xmath1 .
however , demanding this theory to be part of a grand unified theory such as @xmath2 forces the d - parity breaking scale to be very high @xmath3 and hence is in conflict with the constraint from domain wall removal .
we also find another class of r - parity violating susylr models where both these constraints can be simultaneously satisfied . |
the large number of organic compounds and living organisms is based on the unique chemical properties of elemental carbon .
carbon forms a variety of compounds and , in pure form , exists in diamond as well as in graphite structures .
despite these well known properties of carbon , the discovery of fullerenes and nanotubes took the science community by surprise .
the carbon nanotubes were discovered by iijima in 1991 in japan @xcite .
it was soon realized that both semiconducting and metallic nanotubes should exist @xcite .
the mass production of nanotubes succeeded in 1992 @xcite and the first electrical transport measurements in 1996 @xcite .
various possible applications of nanotubes have quickly emerged , for example in molecular electronics.@xcite the coulomb blockade was detected in transport measurements in 1997,@xcite demonstrating that the nanotubes are suitable building blocks of single electron transistors ( sets ) .
the single electron transistor ( set ) has , as an ordinary transistor , three terminals : the source , the drain and the gate .
in addition , the set contains a nearly isolated island between the source and the drain contacts .
the principle of the set is based on the repulsive interaction between electrons which becomes important when the island is smaller than about one micron and its contacts to source and drain are weak ( resistance larger than the quantum of resistance @xmath9 26 k@xmath10 ) .
the first successful set was made by fulton and dolan@xcite at at&t bell laboratories in 1987 using aluminum thin film technology and electron beam lithography . since the early success with the aluminum - set ,
sets have also been manufactured out of two - dimensional electron gas of simosfets , with metal nanoparticles and lately with carbon nanotubes . in case of carbon nanotubes , the nanotube forms the central island and the relatively poor contact resistance between the leads and the nanotube ensures the required isolation . in this paper we present work on basic electronic devices made of mwnts . as a building block of a nanoscale electronic device
, the carbon nanotube may fill many functions . here
we introduce different cases , based on our own measurements , where a multiwalled carbon nanotube is used either as the island of a set , as an ohmic resistor , or as a transmission line .
furthermore , we compare devices made of nanotubes that were synthesized using different methods .
graphite is made out of fairly loosely connected two dimensional carbon layers , each layer having hexagonal lattice structure .
carbon nanotubes are formed from such graphitic layers.@xcite a single walled nanotube ( swnt ) is formed when a piece of graphite sheet is wrapped into a cylindrical form , the edges are seamlessly joined together and the ends of the cylinder are closed . a multiwalled nanotube ( mwnt ) is made out of several concentric graphite cylinders .
the swnts are about 1 - 2 nm in diameter .
the mwnts , in contrast , can be as large as 20 - 30 nm in diameter and are more convenient for experiments .
in constructing a swnt one has several possibilities to cut the original graphite sheet : the width and the length of the sheet can be varied as well as the angle between the symmetry axis of the hexagonal sheet and the main axis of the final cylinder .
the electrical properties of the nanotube are sensitive to the orientation of the hexagonal lattice because it determines the density of electron states at the fermi level .
the band structure of graphite was calculated already in 1947 by wallace using the tight binding approximation@xcite .
a graphite sheet is a semimetal , whose two - dimensional band structure near the fermi surface consists of six conical energy surfaces in the first brillouin zone .
fermi surface of the undoped graphite consists of six points , the vertices of the six cones .
when the two dimensional graphite sheet is wrapped into cylindrical form , the transverse electron motion around the cylinder surface becomes quantized .
the corresponding transverse energy levels of swnts are separated by about 1 ev and only the lowest band is occupied at room temperature and below it .
the electrons on the lowest band move along the cylinder axis and behave truly one - dimensionally .
one can make nanotubes which are several microns long .
the longitudinal quantization of the electron motion leads to the fine structure of the energy bands on the energy scale of 1 mev .
this fine structure is , however , washed away by the broadening of the levels due to impurity scattering .
calculations show that undoped swnts are either semiconductors or metals , approximately one third of the tubes having the metallic orientation between the underlying hexagonal structure and the tube axis . in the semiconducting nanotubes
there are no electron states at the fermi level and the band gap is several hundreds of mevs .
these tubes are good insulators at small bias voltages . in the metallic nanotubes one - dimensional energy bands cross the fermi level .
they are constructed out of the six energy cones of the graphite . because three of the cones are equivalent , the six energy cones of graphite collapse into two independent one - dimensional conduction channels .
taking into account the spin degeneracy of the electrons , the metallic swnts have altogether four independent conduction channels .
the number of conduction channels is important because it determines the electrical conductivity of the nanotube .
the conductivity of a ballistic one - dimensional system is given by @xmath11 , where @xmath12 is the charge of electron , @xmath13 planck s constant and @xmath14 the number of independent conduction channels .
thus the resistance of a ballistic nanotube is expected to be 6.45 k@xmath10 ( @xmath15 ) .
the concentric layers of a mwnt are estimated to be in poor electrical contact to each other .
some experimental results have been interpreted with the assumption that only the outermost layer participates in the electron conduction.@xcite therefore , one expects the electrical properties of the mwnts and swnts to be similar . however
, the metallic inner layers can contribute to the electrical properties of the mwnt at high frequencies.@xcite in fact , the analysis of tunneling experiments in mwnts has recently turned out to be more difficult than expected.@xcite carbon nanotubes are synthesized with various methods that differ from each other in crucial ways with regard to the growth conditions , resulting in nanotubes of different quality . in mwnts in particular , various defects can appear that distort the structure of the ideal mwnt . in the arc - discharge@xcite ( ad ) and
related methods nanotubes are produced in an inert - gas atmosphere from graphite at such a high local temperature that the carbon evaporates and subsequently forms the nanotubes .
later , catalytical synthesis of nanotubes was started using chemical vapor deposition ( cvd ) methods @xcite . in the cvd technique
the nanotubes grow from a catalyst over which a carbon - containing gas is led .
the upper temperature is limited by the requirement that the gas is not decomposed by itself .
therefore the growth temperature with the cvd technique is lower than with the other techniques .
the lower growth temperature of cvd tubes is considered as the main cause for their more defective structure .
ad tubes are generally straight , as shown in fig .
[ sem]a , and exhibit a rather flawless structure in tem images .
indeed , these are rather good approximations of the ideal nanotubes whose basic electronic properties were discussed above .
multiwalled cvd tubes , on the other hand , often exhibit a significant amount of disorder in tem analysis , although images of well ordered sections of individual cvd tubes have also been reported @xcite .
it is likely that cvd material exhibits large variations in quality .
this variation is evident in our experiments on cvd tubes treated with reactive ion etching ( rie ) . in a short etching time
most tubes were uniformly thinned by a factor of 2 , while some tubes had vanished almost completely .
the defects in a low - quality tube are much more susceptible to an etchant than the regular graphitic walls . as a result of the defective structure ,
cvd tubes are generally curved when observed either as freestanding in a tem or as deposited on substrates , as shown in the afm image of fig .
this curvature is rigid and often extends into three dimensions . the nanotube ( diameter 20 nm ) in fig .
[ sem]b has a tail to the left which extends more than 100 nm above the substrate , despite the significant van der waals forces that exist between the substrate and the nanotube .
note that the tail appears broadened due to the interaction between the afm tip and the upwards pointing tail section .
the most impressive of the three dimensionally extended tubes are the regularly coiled 3d spirals of multiwalled cvd - grown tubes @xcite , as shown in fig .
et al._@xcite presented work comparing the elastic moduli of ad and cvd multiwalled tubes , concluding that the former are significantly stiffer than the latter .
thus , although ad tubes have generally better electronic performance from the point of view of device physics , cvd tubes have their own interesting and useful properties .
for electronic transport measurements individual nanotubes have to be connected to external leads .
this is not a trivial task because the nanotubes are quite small , difficult to image and hard to move around . in making simple nanotube circuits
one starts from a solution ( actually a dispersion ) .
a droplet of this solution is placed on the substrate at a desired location , and the solvent is evaporated away .
this leaves behind a number of randomly placed nanotubes on the surface of the substrate .
there are two possibilities to connect the nanotubes to the leads . in the first method
the leads are evaporated on top of the nanotubes .
this method requires accurate imaging of the nanotubes and sophisticated alignment techniques when making the evaporation mask . in the second method one
moves the nanotubes on top of prefabricated gold electrodes ( or makes a large array of electrodes in the hope of finding one nanotube already extending over at least two electrodes ) .
both of these methods are hampered by poor electrical contacts between the nanotubes and electrodes .
this is a problem especially in the case , where nanotubes are deposited on top of the electrodes .
the contact resistances range from a few tens of k@xmath10 to gigaohms and are not well understood .
there are some methods to improve the contacts .
@xcite successfully used electron bombardment in sem , while jeong - o lee _
et al._@xcite used quick heating at high temperature .
we have also used this method in some cases .
contact resistances decrease considerably by a 30 second annealing of the sample in vacuum at 700@xmath16c .
we also have verified the mechanical change in a contact caused by heating : to move the tube after heating required considerably larger applied force .
usually an atomic force microscope ( afm ) is used for imaging surface topography with a tip that has a radius of curvature down to 10 nm .
however , it can also be used to move small objects along the substrates .
we have developed an afm based manipulation method where both moving and imaging are done in the non - contact mode.@xcite in this mode one probes the attractive van der waals force between the vibrating afm tip and the surface . using our method ,
the location of the object can be seen during the movement by monitoring the cantilever oscillation amplitude .
this is how we can build electrical circuits containing several nanotubes , or move nanotubes on top of gold electrodes .
the differences between the arc - discharge and cvd tubes become clearly visible in our afm manipulations of these tubes .
when an ad tube is pushed with an afm tip , it usually either moves as a whole , without bending , or bends sharply at a certain point , with the rest of the tube staying on its place .
these conclusions are similar to those of falvo _ et al . _
@xcite , who studied manipulation of ad tubes more extensively .
we have found that manipulation of those cvd tubes that have significant 3d bending out of the surface is very difficult due to the sticking of the tubes to the afm tip .
also the van der waals forces that hold the nanotube in place and stabilize strained configurations are reduced .
the manipulation of the cvd tubes is also different in the sense that the bending is continuous , without any pivot points , as shown in the manipulation sequence of a cvd tube in fig .
[ manipulation ] .
this is of course expected due to the smaller stiffness of cvd tubes .
thus , only manipulation of ad tubes is convenient , when the required distance is several micrometers .
bending of a nanotube will cause major local changes in the electronic structure , and consequently in intratube transport , as has been shown in calculational work.@xcite these references , however , deal basically with single shell nanotubes . on the experimental side , a large increase in resistance has been observed for swnts upon bending @xcite . for mwnts ,
the effect of bending is quite small compared to the total conductance of the tube , as found by paulson _
et al_.@xcite to observe the effect of bending to the resistance of a nanotube , resistance measurements can be done _ in - situ _ under the afm .
we have , for example , studied one sample , where a mwnt ( of length 1.7 @xmath7 m ) is below the electrodes ( fig .
[ figres ] ) .
the tube is cvd synthesized , as can be seen from its shape . during a long sequence of small movements
, we observed no dramatic changes : the resistance varied between 340 and 380 k@xmath5 , the initial value . both increase and decrease of resistance , typically a few k@xmath5 , was observed between adjacent configurations . based on our experience on cvd tubes
, we expect that a sizable share of the resistance is in the tube itself .
thus we conclude that modest bending of a cvd grown mwnt has only a small effect on the intratube resistance .
the result becomes more general when we note that similar conclusions were obtained@xcite for ad grown mwnts .
it is not surprising that it is more difficult to bend a multiwalled than a single walled tube so sharply that a tunnel junction would emerge at the site of bending .
.basic characteristics of our samples .
the labeling indicates the nanotube synthesis method .
the length @xmath17 refers to the length between the electrodes .
zero - bias resistances are given for two point configuration at @xmath18 k and 4.2 k. the small ratio of @xmath19(4.2 k)/@xmath19(300 k)@xmath20 indicates that our samples are metallic except for ad2 which is semiconducting . [ cols="^,^,^,^,^",options="header " , ]
we have fabricated mwnt - sets both from cvd and ad tubes , some of which are shown in figure [ samples ] .
main characteristics of the samples are given in table 1 .
the nanotubes had typical diameters of 20 nm for the cvd tubes and 15 nm for the ad tubes .
the length of the tube between the electrodes was 0.3 - 5 @xmath21 m for the cvd and 0.3 - 1.7 @xmath21 m for the ad tubes .
two of the cvd tubes were etched in a rie process for 15 seconds which reduced their diameter uniformly from 20 nm to 10 nm .
the electrodes were either fabricated on top of the tubes ( cvd1-cvd3 and ad1 ) or the tubes were placed on top of prefabricated electrodes ( cvd4 , ad2 and ad3 ) .
the electrodes were always made of gold with a chromium sticking layer .
the quality of results obtained on our devices varied over a broad range .
the results on cvd tubes were , in general , more irregular than those on ad tubes .
some of the ad tubes turned out to be excellent islands for sets .
as we will show below , the geometrical structure and the construction method have a clear influence on the results .
figure [ samples](a ) shows an afm image of device cvd1 .
it is made of a 6.8 @xmath7 m long tube which has 25 nm thick gold electrodes evaporated on top of it . at room temperature ,
the resistance of the tube was 200 k@xmath10 .
below 4 k a zero - bias gap due to coulomb blockade became clearly visible .
a few gate modulation curves of this mwnt , measured at @xmath22 k , are shown in fig .
[ cvdmod ] . instead of the expected coulomb blockade period of 1v ( estimated for the employed side - gate
placed a few @xmath21 m away ) , only non - regular gate - modulation traces were found .
this irregular behavior points to the formation of a series of intra - tube quantum dots , caused by disorder .
disorder - induced splitting of a tube into several separate islands has been suggested by mceuen _
et al._@xcite to explain similar experimental results in swnts . from the data set of fig .
[ cvdmod ] and other sets like it , we conclude that cvd tubes , at least with the present amount of disorder , are not very suitable for central islands of set electrometers .
in contrast to cvd - tube devices , sets based on ad tubes display regular @xmath23-curve modulation with respect to the gate voltage . such regularity is particularly nicely illustrated by the results on the device of fig .
[ samples]b ( ad1).@xcite this sample is in fact composed of two crossing mwnts : the lower tube , which acts as the central island of the set , is @xmath24 @xmath7 m long , and the crossing upper tube has been pushed into its position with afm manipulation .
the two - point ( zero - bias ) resistance over the crossing was @xmath25 10 m@xmath10 , which increased to @xmath26 g@xmath10 below 4 k. thus , we could utilize the upper tube for gating the current in the lower tube in this construction .
the device has a room temperature resistance of 71 k@xmath10 .
a coulomb blockade develops fully only at subkelvin temperatures , with a gap of about 1 mv at 150 mk .
figure [ ivcross ] shows the source - drain current @xmath27 as a function of the source - drain bias voltage @xmath28 and the gate voltage @xmath29 , applied to a separate side gate . in general , fourier analysis of the gate modulation curves revealed only one period , indicating the existence of only one island .
in contrast to cvd tubes , we conclude that this tube is not broken into sections , neither by defects nor by the other mwnt placed on top of it .
the gate modulation period with the upper tube as a gate was measured as @xmath30 = 4 mv ( using side gate @xmath30 = 440 mv ) .
the tube had a maximum modulation of 8 na/@xmath12 .
we calculate the gate capacitance to the upper tube as @xmath31 = 40 af ( for side gate @xmath32 = 0.4 af ) . the present configuration with a crossing nanotube as a large - capacitance gate electrode might be useful in certain set applications for reducing cross - talk between the gate and the other electrodes .
furthermore , since the voltage gain of a current - biased set is @xmath33 such a construction allows for devices with high voltage gain @xcite . here
@xmath34 refers to the total island capacitance including the tunnel junctions .
short pieces of nanotubes , equipped with small contact areas , provide a straightforward way to reach sets with large coulomb energies .
figure [ samples]c shows one of the shortest nanotube devices ( ad3 ) that we have made from ad tubes . in our short structures ,
the tubes have been manipulated on top of gold electrodes ( ad2 and ad3 ) .
the length of the tube between the electrodes is 0.3 @xmath21 m for both ad2 and ad3 .
furthermore , the latter is not touching the substrate between the electrodes .
it is separated from the underlying sio@xmath35 by 17 nm .
the room temperature resistance of ad2 was rather high , 5 m@xmath10 , due to weak gold - nanotube contacts , and increased to @xmath25 200 m@xmath10 at 4.2 k. coulomb blockade effects became clearly observable below a few kelvin .
figure [ gatemod ] shows gate modulation of the source - drain current at 4.2 k. the modulation period is 200 mv , giving a gate capacitance of 0.8 af . from measurements of the constant current curves we obtain the junction capacitances and for the charging energy @xmath36 = 2.1 mev .
thus this device is a rather simple implementation of a set working at the relatively high temperature of 4.2 k. moreover , scaling down the dimensions should be straightforward in order to minimize the island size and , consequently , to raise the operating temperature even further .
the low temperature behavior of sample ad2 has been described by roschier _
et al _ @xcite .
nanotubes are quite susceptible to noise caused by charge trapping to surface states.@xcite multiwalled tubes , fortunately , are not as sensitive in this respect as single walled tubes even though some of our mwnt samples do show quite high noise levels .
we obtained the lowest charge noise in device ad3 which had extraordinary properties in several respects when compared with our other samples@xcite . especially , the transport in this tube was close to ballistic , which may be the reason behind its good noise properties .
sample ad3 had a room temperature resistance of 28 k@xmath10 .
at subkelvin temperatures , we measured a total resistance of @xmath37 40 k@xmath10 outside the coulomb blockade regime .
the junction resistance of the nanotube - au contacts are thus less than the quantum of resistance @xmath38 @xmath37 26 k@xmath10 , which means that the coulomb blockade can not fully develop .
consequently , the coulomb oscillations that we measure are smoothened.@xcite as opposed to the usual coulomb blockade , nanotube ad3 exhibited increased conductance around zero bias , which we attribute to resonant tunneling .
only two weakly quantized steps are seen , and therefore this tube can not be said to be fully ballistic .
the ballisticity of freestanding samples is likely to be enhanced most by the decreased capacitive coupling between impurity states of the substrate and the mwnt .
current noise @xmath39 of ad3 , measured at a small voltage bias of 70 @xmath7v , is displayed in fig .
[ noise ] . frequency dependence of the noise power ( @xmath40 ) has a @xmath41 character over the range @xmath42 hz .
the input equivalent charge noise @xmath43 is obtained from the measured current noise according to the formula @xmath44 . in the coulomb blockade regime a modulation of the noise was seen as expected for a set . at a frequency of 45 hz , we obtain the charge noise @xmath45 @xmath1 , which is comparable to the best metallic set devices reported to date @xcite .
theoretically the minimum noise level for a set is @xmath46 , where @xmath47 denotes the frequency range and @xmath48 is the tunneling resistance @xcite .
taking @xmath49 @xmath50 1 and assuming no cotunneling , we obtain the minimum noise as @xmath51 .
this implies that white noise would dominate over @xmath52 noise above 3 khz .
in fact , the shot noise limited region has been reached in our most recent experiments.@xcite
according to simple drude - type arguments , disordered mwnts are good candidates for lumped resistive elements .
as was already mentioned above , we have tried to pursue this idea by selecting a couple of cvd tubes for rie - etching in order to enhance the amount of disorder in them .
the etching halved uniformly the diameter from 20 nm to 10 nm .
gold electrodes were placed on top of these tubes .
the distance between them is 0.3 @xmath7 m .
figure [ cvdrie ] shows the iv characteristics of one of these tubes ( cvd2 ) at low temperatures , compared with an untreated cvd - tube ( cvd4 ) that has roughly the same electrode spacing . at room temperature
we measure a 2-point resistance of 23 k@xmath5 in the rie - etched tube . at @xmath53 mk
we observe only a weak coulomb blockade - type nonlinearity in the iv characteristics and no discernible coulomb oscillations , in contrast with cvd4 .
we therefore assume a relatively small contact resistance and consequently that the measured 2-point resistance indicates the intrinsic resistivity of the rie - etched mwnt to be rather large , on the order of 100 k@xmath5/@xmath7 m .
these results were confirmed in our second rie - etched sample ( cvd3 ) , measured in a 4-point configuration . with this sample we observed , in addition , that the resistance measured between different electrode pairs scales roughly with the distance , in accord with ohmic behavior .
presently we can not tell exactly what role the rie treatment had for the electronic properties of these tubes .
their intrinsic resistance seems to be higher than that of the untreated cvd tubes by a factor @xmath54 .
the high intrinsic resistance together with ohmic addition rule for the nanotube resistance makes the rie - etched cvd tubes simple nanoscale ohmic resistors .
these resistors perform rather well compared with the standard technology utilizing thin chromium films @xcite .
typical thin cr - wires can be employed as linear resistances up to values 10 k@xmath55 m while rie - etched tubes yield 100 k@xmath55 m if their small nonlinearity in the zero - bias region can be accepted . with better metal - tube contacts , this non - linearity can presumably be made even smaller .
in the previous sections ( sects . 4 and 5 ) we treated mwnts as lumped elements of simple electronic circuits .
howewer , their high aspect ratio allows for nanotube lengths up to macroscopic dimensions .
thus any longer section of a mwnt and nearby metallic structures form a transmission line , where the nanotube acts as an inner conductor .
because the electrons move nearly ballistically in a metallic defect - free tube , one naively expects mwnts to be excellent transmission lines for high frequency electromagnetic waves .
however , rather the opposite turns out to be true , since mwnts form special , high impedance transmission lines with a low propagation velocity @xcite . in an ordinary coaxial cable the electromagnetic wave propagates between the metallic outer and inner conductors . the propagation velocity and the impedance of the wave
are determined by the electromagnetic inductance and capacitance , with @xmath56 and @xmath57 .
they are close to the propagation velocity and impedance of electromagnetic waves in free space . in an ordinary coaxial cable ,
the main part of the inductive energy is stored in the magnetic field induced by the moving electrons .
in contrast , in nanotubes the kinetic energy of electrons exceeds their magnetic energy and the inertia of the electrons slows down the electromagnetic waves .
the large kinetic inductance @xmath58 is a consequence of the low electron density of the nanotubes .
the electromagnetic waves traveling along nanotubes are actually electron density waves propagating in one dimension , that is , 1d plasmons .
if the mwnt is regarded as a transmission line , then its impedance will determine the current - voltage characteristics for electron tunneling into the tube over a mesoscopic tunnel junction , as described by the environmental quantum fluctuations theory.@xcite at low voltages , @xmath59 , where @xmath60 . at high voltages ,
the tunnel junction capacitance @xmath61 starts to shunt the environmental impedance @xmath62 , and the power law turns gradually into an ohm s law with a characteristic asymptotic region described by a @xmath63-type of tail.@xcite figure [ taillin ] illustrates the asymptotic approach of the @xmath23-curve of sample ad1 towards ohm s law .
we plot in fig .
[ taillin ] the deficit current @xmath64 which is the deviation of the measured current @xmath27 from the ohmic value given by @xmath65 .
the solid line illustrates the fitting of the theoretical @xmath63-tail : @xmath66 , which yields for this tube a typical value of @xmath67 k@xmath10 . assuming @xmath68 af/@xmath21 m , in fact deduced from the coulomb blockade offset of the same fits , we obtain @xmath69 nh/@xmath7 m .
this inductance value is much higher than expected for magnetic inductance ( @xmath70 ph/@xmath21 m ) but quite well in agreement with our estimate for the kinetic inductance:@xcite @xmath71 nh/@xmath21 m for @xmath15 .
we thus conclude that , due to the large kinetic inductance , the propagation velocity in mwnts is about 1000 km / s and the impedance is on the order of a few k@xmath10 .
carbon nanotubes vary considerably in size and properties . restricting this variation to normal - sized ad and cvd multiwalled nanotubes , we presented in the preceding sections our work on different device configurations that can be realized from mwnts . the difference in structural quality between ad and cvd tubes
was discussed , concluding that ad tubes were preferable from the point of view of device physics .
we have demonstrated the use of multiwalled carbon nanotubes as nanoelectronics building blocks such as sets , resistors and transmission lines .
our set made of two mwnts demonstrated the feasibility of more complex circuits . under proper conditions , such as separating the tube from the substrate , it is possible to minimize the noise level and , in other aspects as well , to approach the theoretical limits of performance . on the other hand , our present understanding of mwnts as transmission lines implies that their kinetic inductance severely limits the propagation speed in them .
nevertheless , the great variety of device structures that can be made from carbon nanotubes suggest that they are among the most promising materials for future molecular electronics .
we acknowledge f. hekking , e. sonin , and a. zaikin for interesting discussions .
this work was supported by the academy of finland and by the large scale installation program ulti - iii of the european union ( hpri-1999-ct-00050 ) . c. schnenberger , a. bachtold , c. strunk , j .-
salvetat , and l. forro , _ applied physics a _ * 69 * , 283 ( 1999 ) ; a. bachtold , c. strunk , j .-
salvetat , j .-
bonard , l. forro , t. nussbaumer , and c. schnenberger , _ nature _ * 397 * , 673 ( 1999 ) .
e.g. _ , c.l . kane and e.j .
mele , _ phys .
lett . _ * 78 * , 1932 ( 1997 ) ; a. rochefort , ph .
avouris , f. lesage , and d.r .
salahub , _ phys .
rev b _ * 60 * , 13824 ( 1999 ) ; m. buongiorno nardelli and j. bernholc , _ phys . rev .
b _ * 60 * , r16338 ( 1999 )
. v.a .
krupenin , d.e .
presnov , m.n .
savvateev , h. scherer , a.b .
zorin , and j. niemeyer , _ j. appl
. phys . _ * 84 * , 3212 ( 1998 ) ; v.a .
krupenin , d.e .
presnov , a.b .
zorin , and j. niemeyer , _
j. low temp .
_ * 118 * , 287 ( 2000 ) .
korotkov , d.v .
averin , k.k .
likharev , and s.a .
vasenko , in _ single - electron tunneling and mesoscopic physics _ , edited by h. koch and h. lubbig ( springer , berlin , 1992 ) , p. 45 ; a. korotkov , _ phys .
b _ * 49 * , 10381 ( 1994 ) . | molecular level components , like carbon multiwalled nanotubes ( mwnt ) , show great potential for future nanoelectronics . at low frequencies ,
only the outermost carbon layer determines the transport properties of the mwnt . due to the multiwalled structure and large capacitive interlayer coupling ,
also the inner layers contribute to the conduction at high frequencies .
consequently , the conduction properties of mwnts are not very far from those of regular conductors with well - defined electrical characteristics . in our work
we have experimentally utilized this fact in constructing various nanoelectronic components out of mwnts , such as single electron transistors ( set ) , lumped resistors , and transmission lines .
we present results on several nano- tube samples , grown both using chemical vapor deposition as well as arc - discharge vaporization .
our results show that set - electrometers with a noise level as low as @xmath0 @xmath1 ( at @xmath2 hz ) can be built using arc - discharge - grown carbon nanotubes
. moreover , short nanotubes with small contact areas are found to work at @xmath3 k with good gate modulation .
reactive ion etching on cvd tubes is employed to produce nearly ohmic components with a resistance of @xmath4 k@xmath5 over a @xmath6 @xmath7 m section . at high frequencies ,
mwnts work over micron distances as special lc - transmission lines with high impedance , on the order of @xmath8 k@xmath5 .
pacs numbers : 73.63.fg , 73.23.-b , 73.23.hk |
although it has been generally accepted that the prompt gamma - ray emission of gamma - ray bursts ( grbs ) results from internal dissipation , likely internal shocks , of a relativistic outflow ( e.g. paczyski & xu 1994 ; rees & 1994 ) , the dissipation site and the radiation mechanism for the gamma - ray emission are still largely unknown .
synchrotron and/or inverse - compton scattering emission by shock - accelerated electrons in the optically - thin region has been proposed as an efficient mechanism for the gamma - ray emission . however , this model does not satisfactorily account for a few observational facts , such as the low - energy spectral slops that are steeper than synchrotron lower energy spectral indices ( preece et al .
2000 ; lloyd et al . 2000 ) , the clustering of peak energies , the correlation between the burst s peak energy and luminosity ( amati et al .
it becomes recognized that an additional thermal component may play a key role and could solve these problems ( e.g. peer et al .
2006 ; ryde et al .
it has also been pointed out that a hybrid model with both a thermal and non - thermal component can describe the spectrum equally well as the band function model ( band 1993 ) , but the former has a more physical meaning ( ryde 2005 ) .
recently , it was suggested that a strong quasi - thermal component could result from the internal dissipation occurring in the optically thick inner parts of relativistic outflows ( rees & 2005 ; peer , & rees 2006 ; thompson et al .
sub - photospheric shock dissipation can increase the radiative efficiency of the outflow , significantly boosting the original thermal photospheric component so that it may well dominate the nonthermal component from optically - thin shocks occurring outside the photosphere .
neutrino emission from gamma - ray bursts has been predicted at different stages of the relativistic outflow , such as the precursor phase ( e.g. bahcall & 2000 ; & waxman 2001 ; razzaque et al .
2003a , b ; razzaque et al . 2004
; ando & beacom 2005 ; horiuchi & ando 2008 ; koers & wijers 2008 ) , the prompt emission phase ( e.g. waxman & bahcall 1997 ; dermer & atoyan 2003 ; guetta et al . 2004 ; murase & nagataki 2006 ; gupta & zhang 2007 ; murase et al .
2006 ) and afterglow phase ( e.g. waxman & bahcall 2001 ; dai & lu 2001 ; dermer 2002 ; li et al . 2002 ; murase & nagataki 2006 ; murase 2007 ; dermer 2007 ) . based on the broken power - law approximation for the spectrum of the prompt emission , presumably from optically - thin internal shocks , a burst of pev neutrinos , produced by photomeson production , was predicted to accompany the prompt gamma - ray emission if protons are present and also accelerated in the shocks ( waxman & bahcall 1997 ) .
the neutrino emission from proton - proton ( @xmath3 ) collisions was generally thought to be negligible due to lower collision opacity for optically - thin internal shocks .
however , as we show below , if some part of the prompt emission arises from internal shocks occurring in the optically - thick inner part of the outflow , as indicated by the thermal emission , a lower energy ( @xmath4 ) neutrino component may appear as a result of @xmath3 collisions .
the photosphere models have been widely discussed in relation to the prompt emission of grbs ( e.g. thompson 1994 ; ghisellini & celotti 1999 ; rees & 2005 ; thompson et al . 2007 ;
ioka et al .
the potential advantage of photosphere models is that the peak energy can be stabilized , which is identified as the thermal or comptonization thermal peak ( see ioka et al .
2007 and references therein ) .
the photosphere radiation may also produce a large number of electron - positron pairs , which may lead to a pair photosphere beyond the baryon - related photosphere ( e.g. rees & 2005 ) , and may also enhance the radiative efficiency ( ioka et al .
2007 ) . on the other hand
, it is also suggested that the number of pairs produced does not exceed the baryon related electrons by a factor larger than a few ( peer et al .
2006 ) . for simplicity
, we here only consider the dissipation below the baryon - related photosphere , which is more favorable for @xmath3 neutrino production .
following rees & ( 2005 ) and peer et al .
( 2006 ) , we assume that during the early stage of the prompt emission , internal shocks of the outflow occur at radii below the baryonic photosphere . initially , the internal energy is released at the base of the outflow , @xmath5 , where @xmath6 and @xmath7 is the schwarzschild radius of a central object of mass @xmath8 .
the internal energy is then converted to the kinetic energy of the flow , whose bulk lorentz factor grows as @xmath9 up to a saturation radius at @xmath10 , where @xmath11 is the initial dimensionless entropy , @xmath12 and @xmath13 are the total energy and mass outflow rates . above the saturation radius ,
the observer - frame photospherical luminosity decreases as @xmath14 and the greater part of energy is in kinetic form , @xmath15 . if the dissipation is maintained all the way to the photosphere , it will lead to an effective luminosity @xmath16 and a temperature ( rees & 2005 ) @xmath17 where @xmath18 is the dissipation efficiency , @xmath19 is the initial temperature of the fireball outflow and @xmath20 is the bulk lorentz factor of the outflow .
the internal shock occurs at @xmath21 , for which the optical depth to thomson scattering by the baryon - related electrons is @xmath22 , where @xmath23 is the thomson scattering cross section and @xmath24 is the kinetic energy luminosity .
the photosphere is further out , at radius @xmath25 .
a detailed calculation taking into account of the electron / positron cooling and the comptonization effect leads to a quasi - thermal emission which peaks at energy @xmath26 for dissipation at thomson optical depth of @xmath27 ( peer et al .
this temperature is consistent with the observed peak energies of prompt gamma - ray emission of a majority of grbs . assuming that a fraction of @xmath28 of the shock internal energy is converted into magnetic fields
, we have a magnetic field @xmath29 . protons accelerated by internal shocks
are assumed to have a spectrum @xmath30 with @xmath31 , as often assumed for non - relativistic or mildly - relativistic shock acceleration .
the maximum proton energy is set by comparing the acceleration time scale @xmath32 with the energy - loss time scales .
the synchrotron loss time is @xmath33 .
assuming that the sub - photosphere emission at the dissipation site peaks at @xmath34 with a thermal - like spectrum , the number density of photons in the comoving frame is @xmath35 .
the @xmath36 cooling time is approximately @xmath37 , where @xmath38 is the inelasticity and @xmath39 is the peak cross section at the @xmath40 resonance . by comparison with the synchrotron loss time
, it is found that the most effective cooling mechanism for protons is the @xmath36 process for protons with energies above the @xmath36 threshold , but below @xmath41 .
equating @xmath42 , we obtain the maximum proton energy in the shock comoving frame @xmath43
the shock - accelerated protons produce mesons via @xmath3 and @xmath36 interactions .
since the meson multiplicity in @xmath3 interactions is about 1 for pions while 0.1 for kaons , neutrinos contributed by pion decay are dominant when the cooling effect of pions is not important , which is applicable to the low energy pp neutrinos .
therefore we here consider only pion production in @xmath3 interactions .
the pion production by @xmath3 interaction in the sub - photosphere dissipation is efficient since the cooling time in the shock comoving frame , @xmath44 can be shorter than the shock dynamic time , @xmath45 , where @xmath46 is the cross section for @xmath3 interactions , @xmath47 is the proton number density , and @xmath48 is the inelasticity .
protons also cool through bethe - heitler interactions ( @xmath49 ) and @xmath36 interactions when the target photon energy seen by the protons is above the threshold energy for each interaction . denoting by @xmath50 the number density of photons in the energy range @xmath51 to @xmath52 , the cooling time in the shock comoving frame for @xmath36 and bethe - heitler cooling processes are given by @xmath53 where @xmath54 , @xmath55 and @xmath56 are respectively the cross section and the inelasticity for @xmath36 ( or bethe - heitler ) process . as a rough estimate ,
the bethe - heitler cooling time is @xmath57 when @xmath58 is a large value ( a good approximation when @xmath59 , chodorowski et al . 1992 ) , where @xmath60 is the thermal peak energy of photons in the comoving frame .
so when the proton energy is larger than @xmath61 , the bethe - heitler cooling dominates over the @xmath3 cooling . at even higher energies near the threshold for @xmath36 interactions at @xmath62 , @xmath36 cooling becomes increasingly dominant .
we compare these three cooling time scales for protons in fig.1 for representative parameters , using more accurate cross sections for @xmath36 and bethe - heitler processes in eq.(4 ) . for photopion production cross section , we take the lorentzian form for the resonance peak ( mcke et al .
2000 ) plus a component contributed by multi - pion production at higher energies , while for bethe - heitler process we use the cross section given by chodorowski et al .
the number density of photons used in the calculation has been assumed to have a blackbody distribution .
the numerical result confirms that that @xmath3 cooling and @xmath36 cooling are dominant , respectively , at lowest and highest energies , while at the intermediate energies , the bethe - heitler cooling is the dominant one .
defining the total cooling time for protons as @xmath63 , the total energy loss fraction of protons is @xmath64 .
the fractions of energy loss by @xmath3 and @xmath36 processes are respectively , @xmath65 the cooling of secondary pions may also affect the neutrino production efficiency if they suffer from cooling before decaying to secondary products .
the pions suffer from radiative cooling due to both synchrotron emission and inverse - compton emission .
the total radiative cooling time is @xmath66\simeq0.002 ( { \epsilon'_{\pi}}/{1 { \rm tev}})^{-1 } \epsilon_{b,-1}^{-1}l_{k,52}^{-1 } r_{11}^2 \gamma_2 ^ 2 \,{\rm s}$ ] , where @xmath67 is the energy density of the magnetic filed in the shock region and @xmath68 is the correction factor accounting for the inverse - compton loss .
the pions also cool due to collisions with protons ( ando & beacom 2005 ) .
the cooling for this hadronic process is @xmath69 , where @xmath70 is the cross section for meson - proton collisions and @xmath71 is the energy lost by the meson per collision .
the suppression of neutrino emission due to cooling of pions can be obtained by comparing the cooling time @xmath72 or @xmath73 with the lifetime of pions @xmath74 in the shock comoving frame , where @xmath75 and @xmath76 are the pion lorentz factor and proper lifetime .
this defines two critical energies for pions , above which the effect of radiative cooling or hadronic cooling starts to suppress the neutrino flux , i.e. @xmath77 and @xmath78 .
the total cooling time of pions is @xmath79 and the total suppression factor on the neutrino flux due to pion cooling is ( razzaque et al .
2004 ) @xmath80
the total energy emitted in neutrinos from @xmath3 or @xmath36 processes per grb is , respectively , @xmath81 where @xmath82 is the energy in accelerated protons in one burst during the dissipative photosphere phase , @xmath83 and @xmath84 are the maximum and minimum energies of acceleration protons . in the absence of pion cooling loss ,
the neutrinos produced by pion decay carry @xmath85 of the energy lost by protons to pion production , since charged and neutral pions are produced with roughly equal probability and muon neutrinos carry roughly @xmath86 of the pion energy in pion decay .
neutrino flux within a factor of 2 ( the factor , however , could be larger for @xmath36 neutrino flux ) . ]
the mean pion energy is about @xmath87 of the energy of the proton producing the pion , so the mean energy of neutrinos is @xmath88 . assuming that protons are efficiently accelerated in shocks with an energy density of @xmath89 , the number of tev neutrinos from one grb is about @xmath90 for @xmath91 , according to eq.(8 ) .
so only from very strong bursts with gamma - ray fluence @xmath92 , which are very rare events , can neutrinos from single grb be detected .
the aggregated muon neutrino flux from all grbs is approximately given by @xmath93 where @xmath94 is the correction factor for the contribution from high redshift sources and @xmath95 is the overall grb rate at redshift @xmath96 . assuming that grb rate traces the star - formation rate in the universe , the calculation gives @xmath97 ( waxman & bahcall 1999 ) .
it is not clear how efficiently the protons are accelerated in grb shocks .
assuming an optimistic case that protons are efficiently accelerated in shocks and that half of the kinetic energy dissipation occurs below the photosphere , we take a mean value @xmath98 for the isotropic equivalent energy in accelerated protons in one grb during the dissipative photosphere phase , based on a typically used value @xmath99 for the isotropic kinetic energy luminosity and a typical long grb duration of @xmath100 s. the grb rate ( e.g. guetta & piran 2007 ; le & dermer 2007 ) , while others get a higher rate comparable to earlier estimate before _ swift _
( e.g. liang et al . 2007 ) . ] at redshift @xmath96 is taken to @xmath101 ( guetta et al .
2005 ; liang et al .
the isotropic luminosity is taken to be @xmath102 .
neutrinos may be reduced if @xmath24 is smaller than @xmath103 . ] the energy - dependent neutrino flux contributed by @xmath3 and @xmath36 interactions are plotted in fig.2 for a set of representative parameters of the dissipative photosphere model and three different dissipation radii .
if the kinetic energy is dissipated at radius of @xmath104 , the calculation ( the red solid curves ) shows that at energies below tens of tev , the neutrino flux is dominated by a @xmath3 component .
taking the detection probability of @xmath105 for tev neutrinos ( gaisser et al .
1995 ) , the expected flux of upward moving muons contributed by this @xmath3 component is about 8 - 10 events each year for a @xmath1 neutrino detector , such as icecube .
we can also estimate the atmospheric neutrino background expectation in coincidence with these grb sources , noting that the search for neutrinos accompanying grbs requires that the neutrinos are coincident in both direction and time with gamma - rays .
taking an average grb duration of @xmath106 s , an angular resolution of icecube of @xmath107 and the atmospheric neutrino background flux of @xmath108 at 1 tev ( ahrens et al .
2004 ) , the atmospheric neutrino background expectation is @xmath109 events from 500 grbs ( in one year ) .
such a low background in coincident with grbs allows the claim of detection of tev neutrinos from grb sources . at energies from a few tev to tens of tev ,
the bethe - heilter cooling suppresses the @xmath3 cooling , resulting in a steepening at several tev in the neutrino spectrum . at energies above the threshold for @xmath36 interactions , the neutrino from @xmath36 process
is heavily suppressed due to the strong radiative cooling of secondary pions . for a larger dissipation radius at @xmath110 ( the blue dashed curves ) , the neutrino emission flux from @xmath36 process
is no longer suppressed and in this case both @xmath3 and @xmath36 neutrino components could be detected by @xmath1 detectors .
we also calculated the neutrino flux , shown by the green dotted lines in fig .
2 , for shock at photosphere radius @xmath111 , assuming that the radiation spectrum is a broken power - law ( due to comptonization ) peaking at @xmath112 , with lower energy and higher energy photon indexes given by @xmath113 and @xmath114 respectively . in this case , the @xmath3 neutrino flux becomes small ( may be marginally detectable ) , while the @xmath36 neutrino flux spectrum is similar to the analytic result obtained by waxman & bahcall ( 1997 ) , as expected for a broken power - law photon spectrum .
note that in one burst the shock dissipation could be continuous from small to large radii , as indicated by the larger variability timescales seen in grbs . by comparing the three cases of different dissipation radii in fig.2
, one can see that as the shock radius increases , the neutrino emission from @xmath3 component decreases , while the @xmath36 component increases until it reaches the saturation level . the total @xmath3 neutrino flux from such continuous dissipation
is thus contributed predominantly by the deepest internal shocks below the photosphere . in the whole neutrino spectrum ,
a valley " is seen between the @xmath3 and @xmath36 components of the spectrum , which may be a potential distinguish feature of the sub - photosphere dissipation effect .
waxman & bahcall ( 1997 ) as well as later works have studied the neutrino emission from the photomeson process during the prompt internal shocks of grbs , assuming that the radiation in the shock region has a broken power - law nonthermal spectrum .
it was found that the neutrino emission peaks at energies above @xmath115 tev . towards lower energies ,
the neutrino emission intensity decreases as @xmath116 ( waxman & bahcall 1997 ) .
however , if internal shocks , especially at the early stage of the prompt emission , occur below the photosphere , a quasi - thermal spectrum will arise . in this _
, we have discussed the neutrino emission associated with the dissipative photosphere that produces such prompt thermal emission .
we find that @xmath3 interaction process becomes important for shock - accelerated protons and provides a new neutrino component , which dominates at energies below tens of tev .
the neutrino emission from photopion process of protons interacting with the sub - photosphere radiation could be significantly suppressed due to radiative cooling of secondary pions , when the dissipation radius is relatively small .
nevertheless , the total contribution by photopion process will not be suppressed since the shock dissipation could be continuous and occur at large radii as well . although tev neutrinos may also be produced during the early precursor stage of grb , i.e. before the jet breaking out the progenitor star ( e.g. razzaque et al .
2004 ; ando & beacom 2005 ) , we want to point out that the tev neutrino component discussed here can be distinguished from them , because in our case the neutrino emission is associated in time with the prompt emission .
after this work has been completed and later put onto the arxiv website ( arxiv:0807.0290 ) , we became aware that k. murase was also working on the sub - photosphere neutrino independently ( k. murase , 2008 , arxiv:0807.0919 ) .
xyw would like to thank p. , s. razzaque , k. murase , e. waxman , z. li and k. ioka for useful comments or discussions .
this work is supported by the national natural science foundation of china under grants 10221001 , 10403002 and 10873009 , the national basic research program of china ( 973 program ) under grants no .
2007cb815404 and 2009cb824800 . | recently , it was suggested that a photospheric component that results from the internal dissipation occurring in the optically thick inner parts of relativistic outflows may be present in the prompt @xmath0/x - ray emission of gamma - ray bursts or x - ray flashes .
we explore high - energy neutrino emission in this dissipative photosphere model , assuming that the composition of the outflow is baryon - dominated .
we find that neutrino emission from proton - proton collision process forms an interesting signature in the neutrino spectra . under favorable conditions for the shock dissipation site
, these low - energy neutrinos could be detected by @xmath1 detectors , such as icecube .
higher energies ( @xmath2 tev ) neutrino emission from proton - proton collision and photo - pion production processes could be significantly suppressed for dissipation at relatively small radii , due to efficient bethe - heitler cooling of protons and/or radiative cooling of the secondary mesons in the photosphere radiation . as the dissipation shocks continue further out
, high energy neutrinos from photo - pion production process becomes dominant . |
efficient simulation of quantum dynamics on classical computers is hampered by the problem of scaling : the complexity of computation in quantum dynamics scales exponentially with the number of degrees of freedom @xcite .
the reason for this exponential growth is the entanglement of the degrees of freedom that is generated during the evolution .
this problem is of a fundamental character : entanglement is viewed as one of the main peculiarities of the quantum dynamics as compared to its classical counterpart @xcite .
asking under what conditions entanglement is generated along the evolution of the quantum system is closely associated with the question of the quantum - classical transition @xcite .
it is customary in quantum dynamical simulations to assume that the initial state of the composite system is factorized in the relevant local basis @xcite .
an important question is whether the product form is conserved along the evolution @xcite .
the answer was generally found to be negative both for the pure @xcite and for the mixed state @xcite dynamics .
a pure composite state is entangled if and only if it is not factorized in the local basis . for mixed states
the situation is more complex @xcite . for a bipartite composite system separability @xcite
is defined as a decomposition of the density matrix of the composite system in the form @xmath5 where @xmath6 and @xmath7 and @xmath8 and @xmath9 are density matrices on hilbert spaces of the first and the second subsystem , respectively .
separable states exhibit only classical correlations .
states that can not be represented in the form ( [ separability ] ) exhibit correlations that can not be explained within any classical theory and are said to be entangled .
there are two qualitatively different kinds of the mixed states entanglement @xcite : free entanglement and bound entanglement .
free entanglement can be brought in a form useful for quantum information processing and bound entanglement is `` useless '' in this sense .
separable states are not of the product form generally . thus the important question remains , under what conditions the mixed state of the composite system evolving from the initial product ( or generally separable ) state develops entanglement along the evolution .
if quantum correlations in the composite system do not develop during the evolution one may speculate that the dynamics of the composite system has classical character .
a possible practical implication is that this `` separable dynamics '' could be simulated efficiently on classical computers .
the dynamics of entanglement was investigated recently in various systems : the quantum brownian particle @xcite , harmonic chain @xcite , two - qubits system interacting with the common harmonic bath @xcite , jaynes - cummings model @xcite , nmr @xcite , various spin systems @xcite , morse oscillator coupled to the spin bath @xcite and bipartite gaussian states in quantum optics @xcite to mention just some cases .
the purpose of the present paper is to explore the temperature dependence of entanglement generation in the course of evolution of a bipartite state in the limit of weak coupling and nonresonant interaction between the parts . under these limitations
nondegenerate perturbation theory was applied to the calculation of the bipartite entanglement in the evolving composite system .
we have considered two cases of interaction - ( 1 ) direct interaction , when two initially disentangled systems are brought into contact ( cf .
fig.[scheme:1 ] ) , and ( 2 ) indirect interaction , when two noninteracting and initially disentangled systems are brought into contact with the third party ( cf .
fig.[scheme:2 ] ) . in each case
the initial state of the composite system was taken to be the product of the thermal states of the parts . to establish quantum entanglement the peres - horodeckii criterion
is employed @xcite .
the peres - horodeckii criterion states that the bipartite system is entangled when the partially transposed density matrix of the system possesses a negative eigenvalue .
the converse statement is generally not true : there exist inseparable states whose partially transposes are positive @xcite .
it is proved in ref.@xcite that states whose partially transposes are positive ( ppt states in what follows ) do not exhibit free entanglement
. therefore ppt states are either separable or bound entangled and as a consequence are not useful in quantum information processing . in the context of simulating a quantum composite system with classical computers , we are interested in the possibility of maintaining a separable form ( cf .
eq.([separability ] ) ) during the evolution .
we conjecture on the basis of ref.@xcite , where it is proved that ppt density matrices of sufficiently small rank are separable , that for a state that remains ppt during the evolution separability can be obtained by embedding in a larger hilbert space . applying the peres - horodeckii criterion to the case ( 1 )
we show that for sufficiently low initial temperature of the subsystems the interaction does induce entanglement unless the ground state of either one of the subsystem is invariant under the interaction . a lower and upper critical temperatures
@xmath3 and @xmath0 exist such that if the composite system evolves from the initial thermal state with temperature @xmath10 the minimal eigenvalue of the partially transposed density matrix becomes negative in the course of the evolution and if @xmath11 the minimal eigenvalue of the partially transposed density matrix stays positive .
the lower bound @xmath2 of the lower critical temperatures @xmath3 was calculated in the limit of weak intersystem coupling and shown to be tight : the negativity of the composite state @xcite , which is a quantitative counterpart of the peres - horodeckii criterion and a measure of entanglement , is shown to be generally negligible for temperatures in the interval @xmath4 . therefore , according to the peres - horodeckii criterion , when @xmath12 the composite system develops entanglement in the course of the evolution and when @xmath13 the composite state remains ppt state .
the question addressed in case ( 2 ) of indirect coupling is what are the conditions on the interaction with the common bath and on the initial temperature of the states which cause entanglement of the noninteracting systems ?
two scenarios with time scales separation are studied : ( a ) two `` slow '' noninteracting systems coupled to a `` fast '' third party ( b ) two `` fast '' noninteracting systems coupled to `` slow '' third party . under some technical assumptions about the form of the interaction we find in both cases that for sufficiently low initial temperature of the noninteracting systems entanglement is induced by the interaction with the third party .
we calculate the lower bound temperature @xmath2 in both cases of the time scales separation . in the system of two noninteracting spins ,
coupled to the common bath , the lower bound coincides with the @xmath0 . in both cases ( 1 ) and ( 2 )
the evolution starts from an uncorrelated initial state of the composite system represented by the tensor product of thermal states of the subsystems involved . as a consequence , initially the eigenstates of the partially transposed density matrix of the composite state are nonnegative .
the evolution under the interaction perturbs the initial state .
the new eigenvalues of the partially transposed density matrix are calculated by the nondegenerate perturbation theory assuming the coupling is weak and the interaction is nonresonanant . the time dependence of the minimal eigenvalue is not analyzed in detail . as the time evolution of the density matrix is quasiperiodic the minimal eigenvalue of the partially transposed density matrix is also a quasiperiodic function of time .
the interaction is said to induce entanglement if the minimal eigenvalue becomes negative in the course of the evolution .
a composite system @xmath14 evolves under the following hamiltonian : @xmath15where @xmath16 , @xmath17 and @xmath18 scales the magnitude of the interaction .
let the initial state be @xmath19 where both @xmath20 and @xmath21 are thermal states : @xmath22 , where @xmath23 is the normalization factor .
the boltzman constant @xmath24 is unity throughout the paper .
the evolution is followed in the interaction picture
. then @xmath25,\end{aligned}\ ] ] where @xmath26 here and in the rest of the paper we take @xmath27 .
it is clear that the density matrix @xmath28 is separable if and only if @xmath29 separable . in
what follows the tags in @xmath29 are omitted for simplicity . in the first order in the coupling @xmath18
the evolution of @xmath30 becomes : @xmath31dt ' .
\label{densityoperator}\end{aligned}\ ] ] entanglment of @xmath32 is established by the application of the peres - horodeckii criterion .
this is carried out by calculating the partial transpose of the state .
the partial transposition @xmath33 with respect to subsystem @xmath34 of a bipartite state @xmath35 expanded in a local orthonormal basis as @xmath36 is defined as : @xmath37 the spectrum of the partially transposed density matrix does not depend on the choice of local basis or on the choice of the subsystem with respect to which the partial transposition is performed . by the peres - horodeckii criterion the eigenvalues of a partially transposed
_ separable _
bipartite state are nonnegative .
the density operator ( [ densityoperator ] ) under the partial transposition ( @xmath38 ) becomes : @xmath39^ { t_a}dt'.\end{aligned}\ ] ] let @xmath40 be the local orthonormal basis of the system @xmath14 composed of the eigenvectors of the unperturbed hamiltonian @xmath41 : @xmath42 where @xmath43 , @xmath44 , is the unperturbed energy spectrum of the hamiltonian @xmath45 .
the initial state is of the tensor product form , cf .
eq.([tensorproduct1 ] ) , therefore : @xmath46 where @xmath47 and @xmath48 are defined by @xmath49 and @xmath50 .
the matrix elements of @xmath51 in the chosen basis are given by : @xmath52 where @xmath53)^ { t_a}\right|j l\right\rangle dt ' \nonumber \\ & = & i\gamma \int_0^t ( \left\langle i k \left| { { \boldsymbol{\mathrm{\hat{\rho}}}}}_a { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t')^t \otimes { { \boldsymbol{\mathrm{\hat{v}}}}}_b(t'){{\boldsymbol{\mathrm{\hat{\rho}}}}}_b \right|j l\right\rangle \nonumber \\ & - & i\gamma \left\langle i k \left| { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t')^t { { \boldsymbol{\mathrm{\hat{\rho}}}}}_a \otimes { { \boldsymbol{\mathrm{\hat{\rho}}}}}_b { { \boldsymbol{\mathrm{\hat{v}}}}}_b(t ' ) \right|j l\right\rangle)dt ' \label{secularmat } \\ & = & i\gamma ( p_{il}-p_{jk})\int_0^t \left\langle i \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_a(t')^t \right|j\right\rangle \left\langle k \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_b(t ' )
\right|l \right\rangle dt'\nonumber \\ & = & i\gamma ( p_{il}-p_{jk})\int_0^t \left\langle j \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_a \right|i \right\rangle \left\langle k \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_b \right|l \right\rangle e^{i t'(e_a^i - e_a^{j}+e_b^l - e_b^{k})}dt'\nonumber \\ & = & \gamma ( p_{il}-p_{jk } ) \left\langle j \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_a \right|i \right\rangle \left\langle k \left|{{\boldsymbol{\mathrm{\hat{v}}}}}_b \right|l\right\rangle \frac{e^{i t(e_a^i - e_a^{j}+e_b^l - e_b^{k})}-1}{(e_a^i - e_a^{j}+e_b^l - e_b^{k } ) } , \nonumber\end{aligned}\ ] ] where @xmath54 designates the transpose of the operator @xmath55 . when @xmath56 , the zero eigenvalue of the initial state @xmath57 is degenerate . as a result
the zero eigenvalue of the partially transposed initial density operator @xmath58 is also degenerate .
the zero eigenvalues correspond to empty initially unoccupied states . by the standard secular perturbation theory the first order correction to the degenerate eigenvalue @xmath59 of the matrix @xmath60
is given by @xmath61 where @xmath62 and @xmath63 are eigenvectors of the matrix @xmath64 , corresponding to the degenerate @xmath59 .
since @xmath65 the eigenvectors of @xmath64 , corresponding to @xmath59 are @xmath66 . therefore at @xmath56 , @xmath67 , and by inspection of eq .
( [ secularmat ] ) , the only nonvanishing matrix elements in the degenerate subspace spanned by @xmath62 and @xmath63 are @xmath68 and @xmath69 where either @xmath70 or @xmath71 .
since the trace of the matrix @xmath72 is zero , either all its eigenvalues vanish or some of them are negative .
all the eigenvalues of @xmath72 can not vanish unless @xmath73 , which from eq .
( [ secularmat ] ) implies @xmath74=0 $ ] or @xmath75=0 $ ] , i.e. the ground state of either one of the subsystem is invariant under the interaction . in this case
the interaction acts locally on the subsystems and can not entangle them .
otherwise there are negative solutions to eq .
( [ eigenproblem ] ) and as a consequence the partial transpose of the density operator attains negative eigenvalues already in the first order in the coupling . therefore , according to the peres - horodeckii criterion , entanglement develops at zero temperature . to simplify the study of the generation of entanglement at finite temperatures it is assumed that the only non zero matrix elements of @xmath76 are those between neighboring states , i.e. @xmath77
. under this assumption
the partially transposed density matrix @xmath51 obtains the following structure : @xmath78 where @xmath79 are defined after the eq.([eigenvectors0 ] ) and @xmath80 by eq.([secularmat ] ) .
there are two kinds of matrix elements @xmath80 : @xmath81 and @xmath82 ( other elements are their counterparts under the transposition ) .
matrix elements @xmath81 couple the unperturbed eigenvalues @xmath83 and @xmath84 . for small coupling strength @xmath18 @xmath85 and the contribution of @xmath81 to
the correction to @xmath83 is negligible and can not make the eigenvalue negative . on the other hand
, the ratio @xmath86 can in general be arbitrary large for low temperatures but for sufficiently high temperatures it tends to zero and as a consequence the contribution of the coupling element @xmath81 to the correction to @xmath87 is negligible .
it can be checked along the same lines that the ratio of the coupling matrix elements @xmath82 to the unperturbed eigenvalues @xmath83 and @xmath88 of the partially transposed density matrix ( [ matrixform ] ) vanish for sufficiently high temperature .
therefore , at least for composite systems with finite hilbert space dimensions , there exists a finite upper critical temperature @xmath0 .
above @xmath0 the spectrum of the partially transposed density matrix remains positive ( ppt ) . in close vicinity of @xmath0 from below the minimal eigenvalue becomes negative in the course of the evolution .
these conclusions stay in accord with a general result @xcite that finite dimensional composite states in sufficiently small neighbourhood of the maximally mixed state ( i.e. thermal states at infinite temperature ) are separable .
we conjecture , that for an infinite composite system , the upper critical temperature @xmath0 exists if the energy spacing is bound . at sufficiently low initial temperature the minimal eigenvalue of
the partially transposed density matrix becomes negative in the course of the evolution .
this means that there exists a finite lower critical temperature @xmath3 .
below @xmath3 the composite systems @xmath89 develops entanglement . in sufficiently close vicinity of @xmath3 from above the state
remains ppt in the course of evolution .
it is possible that @xmath90 .
this equality is confirmed in all numerical tests .
a lower bound @xmath2 for the lower critical temperature @xmath3 can be calculated using perturbation analysis .
it is shown that this bound is tight since the free entanglement in the interval @xmath4 is negligibly small under the weak coupling assumption .
therefore , from the practical point of view the lower bound @xmath2 for @xmath3 can be considered as _ the _ critical temperature for entanglement . for simplicity
the lower bound for the lower critical temperature is termed `` the lower bound temperature '' throughout the paper . at low temperatures the leading order contribution to the negative eigenvalue of the partially transposed density matrix comes from the matrix elements @xmath91 , @xmath92 ( and their complex conjugates ) that do not vanish at @xmath56 .
therefore , to the leading order in @xmath18 , the nonvanishing eigenvalues of the partially transposed density matrix eq.([matrixform ] ) are the eigenvalues of the following effective partially transposed density matrix @xmath93 : @xmath94 the critical temperature , calculated for the effective @xmath95 matrix ( [ matrixform4 ] ) , is a lower bound for the lower critical temperature @xmath3 of the bipartite system @xmath89 .
the eigenvalues of eq .
( [ matrixform4 ] ) are eigenvalues of two @xmath96 matrices : @xmath97 and @xmath98 the eigenvalues of the matrix ( [ twotwo ] ) are : @xmath99 where from eq.([secularmat ] ) : @xmath100 where we define @xmath101 , which is the lowest joint excitation energy of the composite system . from eq .
( [ thelambda ] ) , @xmath102 will be negative whenever @xmath103 and positive if @xmath104 . the lower bound temperature @xmath2 is evaluated from the condition @xmath105 . since @xmath106 is an oscillating function of time ( cf .
( [ eq : a12 ] ) ) the amplitude of @xmath106 is taken to be equal to @xmath107 : @xmath108 assuming that @xmath2 is low @xmath109 and then @xmath110 since @xmath111 is a monotonic function of the temperature , at @xmath13 @xmath112 and at @xmath12 @xmath113 .
finally , the expression for the lower bound temperature @xmath2 becomes : @xmath114 so far only two of the eigenvalues of the matrix ( [ matrixform4 ] ) have been evaluated .
the other two eigenvalues are found to be strictly positive at and above the temperature @xmath2 .
therefore , the expression ( [ temperature ] ) defines the critical temperature for the partially transposed effective density matrix ( [ matrixform4 ] ) and the lower bound temperature of the partially transposed density matrix ( [ matrixform ] ) .
( [ temperature ] ) can be generalized to an interaction term of the form @xmath115 : @xmath116 provided @xmath117 .
when this term vanishes there is no entanglement in the first order in the coupling strength @xmath18 . for the system of two interacting spins the lower bound @xmath2 given by eq .
( [ temperature ] ) coincides with the upper critical temperature @xmath0 therefore in this case _ the _ critical temperature exists in the strict sense .
[ fig:1 ] shows results of the numerical calculation of the critical temperature as a function of coupling strength for a system of two interacting spins evolving from the initial product state of two thermal states .
the peres - horodeckii criterion was used and the partial transpose of the evolving density matrix was calculated numerically to determine entanglement .
the shaded area in the parametric space of the logarithm of inverse coupling and the inverse initial temperature represents the values of the parameters where no entanglement develops .
for coupling up to @xmath118 @xmath2 given by eq .
( [ temperature ] ) ( the dashed line ) corresponds well to the numerical values of @xmath0 .
it is interesting to note that for large values of coupling the critical temperature asymptotically tends to a finite constant value of the same order of magnitude as the energy difference between the first excited and the ground state of the unperturbed composite system . at @xmath12 the minimal eigenvalue of the partially transposed state ( [ matrixform ] )
is negative .
we want to show that above @xmath2 the negative eigenvalues of the matrix ( [ matrixform ] ) are of higher order in @xmath18 and therefore are negligibly small when the coupling is weak .
let s consider corrections to the eigenvalues @xmath119 and @xmath120 of the composite state ( [ matrixform ] ) .
the order of magnitude estimate of the smallest one of the corrected eigenvalues is : @xmath121 , where @xmath122 . for simplicity
we assume @xmath123
. then @xmath124 .
below @xmath2 the minimal eigenvalue of the state ( [ matrixform ] ) is @xmath125 .
we shall estimate the ratio @xmath126 and show that it is negligible when the coupling is weak .
we shall assume without loss of generality that the ground state energy is zero : @xmath127 .
then the partition function @xmath128 of the composite system is larger than unity .
it follows that @xmath129 we are looking for the maximal value of @xmath130 in the interval @xmath131 , corresponding to the condition @xmath132 , i.e. to the negative values of @xmath133 .
@xmath134 is determined by the condition @xmath135 .
the ratio @xmath130 is positive in the interval @xmath131 and vanishes on its borders .
therefore @xmath130 has a maximum @xmath136 at @xmath137 , which is found from the condition @xmath138 .
the calculation gives @xmath139 , which proves that there is one maximum @xmath136 at @xmath137 .
we remark , that @xmath134 , corresponding to the largest @xmath140 over all @xmath141 and @xmath142 , @xmath143 , is of the order of the upper critial temperature @xmath144 .
the maximal value of @xmath130 is given by : @xmath145 where the inequality follows from the fact that @xmath146 in general .
as a next step we notice that @xmath147 , therefore @xmath148 introducing the definition @xmath149 and taking @xmath150 , which corresponds to a rescaling of the coupling strength @xmath18 , leads to : @xmath151 typically the spectrum becomes denser with increasing energy . in that case @xmath152 .
values of @xmath133 , corresponding to @xmath153 need not be taken into account , because @xmath154 in this case and as a conseqence @xmath155 at @xmath156 .
at @xmath157 the upper bound for @xmath158 scales as @xmath159 and therefore corresponding negative eigenvalues of eq .
( [ matrixform ] ) are negligible . in this case
we expect that @xmath160 . in those cases
when @xmath161 the upper bound for @xmath158 scales as @xmath162 and the corresponding negative eigenvalues of eq .
( [ matrixform ] ) can be neglected , too .
when @xmath163 is moderately larger than unity the upper bound eq.([scaling ] ) for @xmath130 has a local maximum .
the position of the maximum weakly depends on @xmath18 : numerical calculations show @xmath164 in the range of @xmath165 .
the value of the minimum is a monotonically slowly increasing function of @xmath18 . in the range @xmath165 numerical estimation of eq.([scaling ] )
shows values @xmath166 for the local maximum .
it is clear that the upper bound eq.([scaling ] ) for @xmath130 is far from being tight .
in fact , numerical calculations show that @xmath130 is generally much smaller . as a consequence , the corresponding negative eigenvalues of eq .
( [ matrixform ] ) can be neglected .
it can be argued that although each one of the negative eigenvalues of eq .
( [ matrixform ] ) is negligible at @xmath156 the ( free ) entanglement of the state can not be neglected .
in fact , the minimal negative eigenvalue of the partially transposed matrix is not a measure of entanglement .
various measures of entanglement have been defined @xcite . in the present context
we will employ a quantitative counterpart of the peres - horodeckii criterion , the negativity @xcite : @xmath167 where @xmath168 is the trace norm of an operator @xmath55 .
the negativity of the state equals the absolute value of the sum of the negative eigenvalues of the partially transposed state .
when the negativity of a composite bipartite state vanishes there is no free entanglement in the state .
it can be shown by the order of magnitude analysis similar to the analysis above that values of the negativity of the composite state , corresponding to the partial transpose ( [ matrixform ] ) , are generally dominated by the minimal negative eigenvalue . as a consequence , the negativity of the state , evolving from the initial thermal product state at the temperature @xmath156 , is negligible under the weak coupling assumption .
figures [ fig:7 ] and [ fig:6 ] display results of numerical calculations of the time averaged negativity of the composite state ( [ densityoperator ] ) as a function of initial temperature for two different kinds of unperturbed spectra of the composite system @xmath89 . both @xmath34 and @xmath169
are four level systems .
the composite system evolves from the initial product of thermal states of @xmath34 and @xmath169 under the hamiltonian ( [ hamiltonian1 ] ) .
[ fig:7 ] presents the results of calculations for the following choice of the unperturbed spectra of @xmath170 and @xmath171 : @xmath172 and @xmath173 .
care was taken to avoid resonances and the spectra were chosen to become denser with increasing energy .
the interaction terms in the hamiltonian were restricted to @xmath174 and the coupling strength @xmath175 .
we see that @xmath176 and the time averaged negativity @xmath177 is negligible in the interval @xmath4 as expected .
[ fig:6 ] displays the time averaged negativity @xmath177 as a function of initial temperature of the composite state of two interacting four level subsystems @xmath34 and @xmath169 with the unperturbed energy spectra @xmath178 and @xmath173 .
the composite state evolves from the initial product of two thermal states under the hamiltonian ( [ hamiltonian1 ] ) , where @xmath174 and the coupling strength @xmath175 . in choosing the unperturbed spectra care was taken to avoid resonances and to ensure that the maximal value of @xmath149 equals the position of the local maximum of the upper bound ( [ scaling ] ) , corresponding to @xmath175 .
[ fig:7 ] shows that the time averaged negativity @xmath177 is negligible in the interval @xmath4 as expected .
the value of @xmath179 ( the definition of @xmath134 is given after eq.([inequality ] ) ) , corresponding to the maximal value @xmath180 is calculated .
@xmath143 is in good correspondence with the value @xmath0 , calculated numerically .
the dynamics studied is of the composite system @xmath181 where systems @xmath169 and @xmath182 do not interact directly ( cf .
fig.[scheme:2 ] ) .
the entanglement explored is of the reduced composite system @xmath183 .
the evolution is generated by the following hamiltonian : @xmath184 where @xmath185 and @xmath186 .
the analysis is carried out in the interaction picture .
the initial state is taken to be @xmath187 , where @xmath20 , @xmath21 and @xmath188 are thermal states .
since @xmath169 and @xmath182 are noninteracting entanglement will appear only in the second order in the coupling .
up to second order in @xmath18 the state of the composite system @xmath181 becomes : @xmath189dt'\nonumber \\ & -&\gamma^2 \int_0^t \int_0^{t'}\left[{{\boldsymbol{\mathrm{\hat{v}}}}}(t'),\left[{{\boldsymbol{\mathrm{\hat{v}}}}}(t '' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}(0)'\right ] \right]dt'dt'',\end{aligned}\ ] ] where @xmath190 in what follows the tag above the @xmath32 is omitted .
next the system is reduced to @xmath183 by taking the partial trace of @xmath32 over the system @xmath34 degrees of freedom and the partial transposition with respect to the subsystem @xmath169 is taken : @xmath191 where @xmath192 and @xmath193\right)^ { t_b}dt'\nonumber \\ & -&\gamma^2 \int_0^t \int_0^{t'}tr_a \left ( \left[{{\boldsymbol{\mathrm{\hat{v}}}}}(t'),\left[{{\boldsymbol{\mathrm{\hat{v}}}}}(t '' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}(0)\right ] \right ] \right)^ { t_b}dt'dt''.\label{secondorderdensity}\end{aligned}\ ] ]
let @xmath40 be the local orthonormal basis of the system @xmath194 composed of the eigenstates of the hamiltonian @xmath195 : @xmath196 where @xmath197 , @xmath44 , is the unperturbed energy spectrum of the hamiltonian @xmath198 .
since @xmath199 : @xmath200 where @xmath201 , and @xmath48 are defined by @xmath202 and @xmath203 .
the matrix elements of @xmath204 are given by : @xmath205 where by definition @xmath206 .
from this point the calculations proceed along the same lines as in section ii following eq.([rhoelements ] ) .
the minimal eigenvalue of the partially transposed reduced state @xmath204 is shown to be negative at sufficiently low temperatures and the lower bound temperature @xmath2 is calculated .
the negative eigenvalue of the partially transposed composite state eq.([rhobc ] ) is calculated to the leading order in the coupling strength @xmath18 assuming @xmath207 . as in section ii
the eigenvalue is found from the spectrum of the @xmath208 matrix : @xmath209 completely analogous to the matrix ( [ twotwo ] ) .
the eigenvalues of eq .
( [ twotwo2 ] ) are : @xmath210 and the eigenvalue @xmath102 becomes negative when @xmath211 . to calculate @xmath212 , @xmath213 and @xmath92
we first note that the integrand in the first order term in eq .
( [ secondorderdensity ] ) is : @xmath214\right)^{t_b}= \left\langle { { \boldsymbol{\mathrm{\hat{v}}}}}_a\right\rangle \left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t'),{{\boldsymbol{\mathrm{\hat{\rho}}}}}_{bc}(0 ) \right]^{t_b } \nonumber \\ & = & \left\langle { { \boldsymbol{\mathrm{\hat{v}}}}}_a\right\rangle \left(\left[{{\boldsymbol{\mathrm{\hat{v}}}}}_b(t ' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}_b\right]^t\otimes { { \boldsymbol{\mathrm{\hat{\rho}}}}}_c + { { \boldsymbol{\mathrm{\hat{\rho}}}}}_b^t\otimes \left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_c(t'),{{\boldsymbol{\mathrm{\hat{\rho}}}}}_c \right ] \right ) \label{firstorder } \\ & = & -\left\langle { { \boldsymbol{\mathrm{\hat{v}}}}}_a\right\rangle\left(\left[{{\boldsymbol{\mathrm{\hat{v}}}}}_b(t')^t , { { \boldsymbol{\mathrm{\hat{\rho}}}}}_b\right]\otimes { { \boldsymbol{\mathrm{\hat{\rho}}}}}_c - { { \boldsymbol{\mathrm{\hat{\rho}}}}}_b\otimes \left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_c(t'),{{\boldsymbol{\mathrm{\hat{\rho}}}}}_c \right]\right),\nonumber \end{aligned}\ ] ] where @xmath215 means the thermal average of the operator @xmath216 and the notation @xmath217 is introduced .
the initial condition @xmath218 was used . since @xmath219 the term eq .
( [ firstorder ] ) does not contribute to the eigenvalues of the matrix ( [ twotwo2 ] ) in the first order . to simplify the calculation of the second order corrections it is assumed that the thermal average of the system @xmath34 coupling operator @xmath215 vanishes .
this assumption is not crucial for the qualitative picture of temperature dependence of the entanglement .
moreover , it is in line with common models of coupling , for example , the caldeira - leggett model @xcite , dipole interaction with the electromagnetic field @xcite , etc . the integrand in the second order term in eq . ( [ secondorderdensity ] ) is : @xmath220 \right]\right)^{t_b}=\left\langle { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t ' ) { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t'')\right\rangle \left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t'),{{\boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t '' ) { { \boldsymbol{\mathrm{\hat{\rho}}}}}_{bc } \right]^{t_b}\nonumber \\ & -&\left\langle { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t '' ) { { \boldsymbol{\mathrm{\hat{v}}}}}_a(t')\right\rangle \left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t ' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}_{bc } { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t'')\right]^{t_b}.\end{aligned}\ ] ] expanding the thermal averages in the orthonormal basis @xmath221 of the hamiltonian @xmath222 leads to : ( 3.375,0 ) @xmath223 \right]\right)^{t_b } & = & \sum_{m , n}p_{a , n}\left|\left\langle m\left| { { \boldsymbol{\mathrm{\hat{v}}}}}_a\right|n\right\rangle\right|^2 ( \cos(\omega_{m n}^a(t'-t''))\left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t'),\left[{{\boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t '' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}_{bc } \right]\right]\nonumber \\ & + & i\sin(\omega_{m n}^a(t'-t''))\left [ { { \boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t'),\left\{{{\boldsymbol{\mathrm{\hat{v}}}}}_{bc}(t '' ) , { { \boldsymbol{\mathrm{\hat{\rho}}}}}_{bc } \right\}\right])^{t_b } \label{secondorder},\end{aligned}\ ] ] ( 3.375,0 ) ( 0,0)(1,0)3.375 ( 0,0)(0,-1)0.08 where @xmath224 is the energy difference between the states @xmath221 and @xmath225 of the hamiltonian @xmath222 , the @xmath226 designates anticommutator of operators @xmath55 and @xmath227 and @xmath228 . for simplicity
the notation @xmath229 is used for the operator ( [ secondorder ] ) . expressing the operator @xmath230 in terms of @xmath231 and
@xmath232 we put the matrix elements of @xmath229 into the following form : ( 3.375,0 ) @xmath233 ( 3.375,0 ) ( 0,0)(1,0)3.375 ( 0,0)(0,-1)0.08 where @xmath234 stands for the energy difference between the first excited and the ground states of the unperturbed subsystem @xmath169 ( @xmath182 ) .
the matrix elements @xmath212 , @xmath213 and @xmath92 are given by : @xmath235 the integration is straightforward but the final expressions are cumbersome .
two cases are considered explicitly : @xmath236 @xmath237 and @xmath238 @xmath239 . in both cases
it is shown that at sufficiently low initial temperature of the system @xmath240 one of the eigenvalues of the matrix ( [ twotwo2 ] ) is negative and the lower bound temperature @xmath2 is calculated .
performing the integrations in eq .
( [ matrix ] ) and taking the leading terms in @xmath241 brings to : ( 3.375,0 ) @xmath242 ( 3.375,0 ) ( 0,0)(1,0)3.375 ( 0,0)(0,-1)0.08 at @xmath56 the minimal eigenvalue of eq .
( [ thenextlambda ] ) is given by @xmath243 , which to the leading order in @xmath241 gives @xmath244 .
this proves that the system @xmath183 becomes entangled at sufficiently low temperature .
we note that this result holds at any finite temperature of the system @xmath34 . at infinite temperature of the system
@xmath34 @xmath245 and no free entanglement is generated in the system @xmath183 . at finite initial temperature of @xmath183 the condition @xmath113 translates to @xmath246 to the leading order in @xmath247 .
the lower bound temperature @xmath2 is found from the condition @xmath248 . since @xmath106 is an oscillating function of time the amplitude of @xmath106
must be substituted for @xmath106 in this equality , which leads to the following equation defining the lower bound temperature : @xmath249 finally leading to : @xmath250 a generalization of the formula to the case of interaction of the form @xmath251 can be carried out along the same lines .
the entanglement in the reduced system of two noninteracting `` slow '' spins interacting with the `` fast '' four level `` bath '' was explored numerically and the results are plotted on fig .
[ fig:2 ] .
the shaded area in the parametric space of the logarithm of inverse coupling strength and the inverse initial temperature of the spins represents parametric values for which no entanglement develops in the course of the evolution .
the border of the shaded area corresponds to the critical temperature for various coupling magnitudes .
the hamiltonian of the composite system is : @xmath252 where @xmath253 , @xmath254 and @xmath255 .
the temperature of the thermal initial state of the `` bath '' is @xmath256 .
the value of @xmath257 chosen for the numerical calculation is unity .
the correspondence of eq .
( [ temperature2 ] ) ( the dashed line ) to the numerical values is very good up to a coupling strength of the order of unity .
we note that for large values of the coupling strength @xmath18 the critical temperature asymptotically tends to a finite constant value .
the case @xmath239 is more complex . to demonstrate entanglement at zero temperature of the system
@xmath183 two simplifying assumptions were added .
the first is that the temperature of the system @xmath34 is also zero .
the second is that the matrix elements of @xmath258 couple only the neighboring states : @xmath259 . under these two assumptions
the expressions for @xmath212 , @xmath213 and @xmath260 become : @xmath261 where @xmath262.\end{aligned}\ ] ] to estimate @xmath263 new variables @xmath264 , @xmath265 and @xmath266 are introduced . ignoring the zero measure set of commensurable frequencies we can treat the function @xmath263 as function of three independent variables @xmath267 , @xmath268 and @xmath269 .
the range of @xmath263 in the cube , defined by @xmath270 , can be explored numerically and is found to be : @xmath271 , where @xmath272 .
therefore , from eq.([expression ] ) @xmath273 , which proves that at zero temperature @xmath113 ( cf .
eq.([thenextlambda ] ) ) and the systems @xmath169 and @xmath182 are entangled by the interaction with the system @xmath34 .
the lower bound temperature is determined by the condition @xmath274 , which translates to @xmath275 ( cf .
eq.([thenextlambda ] ) ) . the latter condition can be put in the form @xmath276 . since @xmath92 and @xmath213 are nonnegative independent functions of time the minimum value of @xmath277
is obtained at @xmath278 .
then the lower bound temperature can be calculated from the condition that the amplitude of @xmath279 equals @xmath280 : @xmath281 finally leading to : @xmath282 it is interesting to note that @xmath2 in this case does not depend on the time scales of the `` slow '' system . the entanglement in the reduced system of two noninteracting `` fast '' spins interacting with the `` slow '' four level `` bath '' was explored numerically and the results are plotted on fig .
[ fig:3 ] .
the shaded area in the parametric space of the logarithm of inverse coupling strength and the inverse initial temperature of the spins represents parametric values for which no entanglement develops in the course of the evolution .
the border of the shaded area corresponds to the critical temperature for various coupling magnitudes .
the hamiltonian is chosen to be similar to the previous example , cf .
eq.([figure2 ] ) , but time scales of the subsystems are reversed : @xmath283 where @xmath253 , @xmath284 and @xmath255 .
the temperature of the thermal initial state of the `` bath '' was chosen as @xmath285 , which is small compared to the energy scale of the `` bath '' chosen for the numerical calculation : @xmath286 . the dashed line on the fig .
[ fig:3 ] is a plot of eq .
( [ temperature3 ] ) and the correspondence to the border of the shaded area at coupling strength up to the order of unity is good .
entanglement is created by both direct and indirect weak interaction between two initially disentangled systems prepared in thermal states at sufficiently low temperatures .
the study is restricted to the conditions where the ground states of both systems are not invariant under the interaction and the interaction is nonresonant . as a consequence ,
the present analysis left out some interesting models such as the jaynes - cummings model @xcite .
the jaynes - cummings model of interacting two level system and a quantized field mode was investigated in ref.@xcite .
it was found that no free entanglement is generated in the course of the evolution of the composite system if the initial temperature of both the subsystems is sufficiently high . in the case of indirect interaction to show entanglement at @xmath56 we have assumed that the thermal average of the third party coupling term in the initial state vanishes .
the reason for the assumption was technical .
it should be noted that many system - bath models of linear coupling satisfy this assumption @xcite .
the additional technical assumption was that the coupling terms of the noninteracting parties possess matrix elements only between the adjacent energy states . here , too , the assumption is general for weak coupling models .
two cases of time scale separation were considered explicitly .
the first is the case of two `` slow '' systems interacting via the `` fast '' third common party .
the second is the case of two `` fast '' systems interacting via the `` slow '' third common party . in the first case
the entanglement was shown to appear at sufficiently low initial temperature of the `` slow '' systems for any finite temperature of the third party . in the second case the entanglement develops at sufficiently low initial temperature of the `` fast '' systems . in this case
we assumed that the third party was prepared at zero temperature and that the third party coupling agent has nonvanishing matrix elements only between the adjacent energy states .
this assumption is stronger than just assuming that its thermal average vanishes . in these cases of indirect interaction and in the case of the direct interaction between the parts
we have shown that if the initial temperature of the bipartite state is zero entanglement is generated by the interaction . at sufficiently high temperature
the composite state remains ppt in the course of evolution . from these results
it follows that a lower critical temperature @xmath3 exists : if the initial temperature of both thermal states is below @xmath3 the interaction generates entanglement in the course of the evolution , and if the initial temperature is sufficiently close to @xmath3 from above the the composite state remains ppt forever .
when the composite system is finite dimensional there exists an upper critical temperature @xmath0 : if the initial temperature of both thermal states is higher than @xmath0 the composite state remains ppt in the course of evolution and if the initial temperature is sufficiently close to @xmath0 from below entanglement is generated .
we conjecture on the basis of numerical experiments that @xmath287 in general . in both cases of a direct and an indirect interaction between the initially disentangled systems , prepared in thermal states , we calculated the lower bound @xmath2 for the lower critical temperature @xmath3 .
when the initial temperature of both thermal states is below @xmath2 the interaction generates entanglement in the course of the evolution . for temperatures above the lower bound @xmath2
the negativity of the partially transposed composite state is zero in the leading order in the coupling strength and therefore negligible in the weak coupling limit .
it follows , that @xmath2 may be considered as the _ physical _ critical temperature for the negativity of the composite state .
separable states can be considered as classical states , because they lack quantum correlations .
one may hope that , as a consequence , the dynamics of separable states can be efficiently simulated on classical computers . whether this is possible
is an open question in quantum information science .
if a moderate scaling procedure exists for the simulation of the dynamics of a separable bipartite state , then it seems that such a procedure exists also if the evolving state remains ppt for all times .
ref.@xcite has proved that a density operator @xmath30 supported on a @xmath288 dimensional hilbert space ( @xmath289 ) with positive partial transpose and a rank smaller than or equal to @xmath290 is separable .
it follows that a ppt state of dimension @xmath290 is always separable when embedded in the larger hilbert space of dimension @xmath291 or higher .
the dynamics of the low dimensional ppt state will be physically equivalent to the dynamics of the high dimensional separable state which can ( hopefully ) be simulated efficiently on the classical computer .
the present analysis shows that above a critical temperature @xmath2 the ppt character of a composite state is preserved along the evolution .
the challenge is to construct an effective simulation for the dynamics of a composite quantum systems at finite temperature employing classically based computers .
we want to dedicate this study to assher peres who passed away on february 2005 .
we are grateful to roi baer and jose palao for critical comments .
this work is supported by dip and the israel science foundation .
r. feynman , int .
, * 21 * , 467 ( 1982 ) .
e. schroedinger , naturwissenschaften * 23 * , 807 , ( 1935 ) ; * 23*,823,(1935 ) ; * 23*,844,(1935 ) .
a. peres , _ quantum theory : concepts and methods _ kluwer academic publishers ( dordrecht , nerelands1998 ) .
zurek , rev .
phys . * 75 * 715(2003 ) .
see d. giulini , e .
kiefer , j.kupsch , i.o .
stamatescu and h.d .
zeh ( eds . ) , _ decoherence and the appearance of a classical world in quantum theory_,2nd edition , ( springer , berlin , heidelberg , new - york , 2003 ) .
e. b. davis , rep .
math . phys . * 11 * , 169 ( 1977 ) .
g. lindblad , j. phys .
a * 29*,4197(1996 ) .
j. gemmer and g. mahler , euro.phys.j .
d * * 17**385 ( 2001 ) .
t. durt , z. naturforsch .
a * 50 * , 425 ( 2004 ) .
d.bruss , j. math . phys . * 43 * , 4237 ( 2002 ) .
r. f. werner , phys .
, a * 40 * , 4277 ( 1989 ) .
m. horodecki , p.horodecki and r.horodecki , phys .
lett . * 80 * , 5239 ( 1998 ) .
j. eisert and m.b .
plenio , phys .
* 89 * 137902 ( 2002 ) .
k. audenaert , j.eisert,m.b .
plenio and r.f .
werner , phys .
a * 66 * 042327 ( 2002 ) .
d. braun , phys .
89 , 277901 ( 2002 ) .
s. scheel , j. eisert , p. l. knight , and m. b. plenio , j. mod .
opt . * 50 * , 881 ( 2003 ) .
doronin , phys . rev . a * 68 * 052306 ( 2003 ) .
a. hutton and s. bose , e - print quant - ph/0408077 ( 2004 ) .
a. hutton and s. bose , phys .
a * 69 * 042312(7 ) ( 2004 ) .
a. sen ( de ) , u. sen , m. lewenstein , e - print quant - ph/0505006 ( 2005 ) .
d. gelman , c.p .
koch and r. kosloff , j. chem .
phys . * 121 * 661 ( 2004 ) .
dodonov , v.v .
dodonov and s.s .
mizrahi , j. phys .
a * 38 * 683 ( 2005 ) .
a. peres , phys .
lett . * 77 * 1413 ( 1996 ) .
m. horodecki , p.horodecki and r.horodecki , phys .
a**223 * * , 1 ( 1996 ) .
p. horodecki , phys .
a * 232 * , 33 ( 1997 ) .
p. horodecki , m.lewenstein , g. vidal and i.cirac , phys .
a**62 * * , 032310 ( 2000 ) .
g. vidal , r. f. werner , phys .
, a * 65 * , 032314 ( 2002 ) .
l. gurvits and h. barnum , phys .
a * 66 * 062311 ( 2002 ) .
s. bandyopadhyay and v. roychowdhury , phys .
a * 69 * 040302(r ) ( 2004 ) .
m. plenio , s. virmani , e - print quant - ph/0504163 ( 2005 ) .
a. o. caldeira and a. j. leggett , physics a * 121 * , 587 ( 1983 ) ; h. carmichael , _ an open system approach to quantum optics _ , ( springer - verlag , berlin , heidelberg , new - york , 1993 ) .
jaynes and f.w .
cummings , proc .
ieee * 51 * , 89 ( 1963 ) .
m. khasin and r. kosloff , in preparation . | both direct and indirect weak nonresonant interactions are shown to produce entanglement between two initially disentangled systems prepared as a tensor product of thermal states , provided the initial temperature is sufficiently low .
entanglement is determined by the peres - horodeckii criterion , which establishes that a composite state is entangled if its partial transpose is not positive .
if the initial temperature of the thermal states is higher than an upper critical value @xmath0 the minimal eigenvalue of the partially transposed density matrix of the composite state remains positive in the course of the evolution .
if the initial temperature of the thermal states is lower than a lower critical value @xmath1 the minimal eigenvalue of the partially transposed density matrix of the composite state becomes negative which means that entanglement develops .
we calculate the lower bound @xmath2 for @xmath3 and show that the negativity of the composite state is negligibly small in the interval @xmath4 . therefore the lower bound temperature @xmath2 can be considered as _ the _ critical temperature for the generation of entanglement . |
two - particle irreducible ( 2pi ) effective actions and approximation schemes based on them are often touted as useful techniques when it is necessary to go beyond the standard perturbative field theory , with applications to thermal and non - equilibrium plasmas / fluids , strongly coupled quantum field theories and systems dominated by many - body collective effects .
the technique was originally developed by lee and yang @xcite , luttinger and ward @xcite , baym @xcite and others in the context of many - body theory , then extended by cornwall , jackiw and tomboulis @xcite to relativistic field theory where it found its natural formulation in terms of functional integrals . since
then a broad literature has developed surrounding 2pi effective actions and their generalizations ( see @xcite for a good introductory review ) .
approximations based on 2pi effective actions are often justified as a selective re - summation of perturbation theory ( some recent examples : @xcite ; interestingly though , this point of view is not found in the cornwall , jackiw and tomboulis paper @xcite ) , however they are not really a summation method in the same sense as , for example , borel summation . rather , this interpretation comes in a rather roundabout way .
first , a set of self - consistent equations of motion is derived for the mean field @xmath1 and connected correlation function @xmath2 , starting directly from the non - perturbative definition of the theory through the path integral . only after formally solving the equations of motion by repeatedly iterating them
does one obtain the usual perturbative expansion for these quantities or , if the equations of motion are truncated , a selective re - summation of perturbation theory appears . in this work we attempt to clarify the connection between 2pi approximations , traditional perturbation theory and re - summation methods with a special emphasis on what analytic features of the 2pi formalism allow improved approximations to be obtained in the presence of a divergent feynman diagram series .
unfortunately robust comparisons are difficult because the large order behaviour of perturbation theory is known , at best , in a sketchy form for most field theories of interest . to that end
we restrict attention to a genuinely trivial model `` field theory '' in zero spacetime dimensions ( i.e. probability theory ) for which exact results are easily obtainable and all complications due to renormalization etc .
this model nevertheless accurately represents the typical combinatoric structure of large order perturbation theory , at least in those cases where the behaviour is known in more interesting field theories .
we also introduce a `` spectral function '' representation of the green function ( similar to the one first introduced by bender and wu @xcite ) to capture the non - analyticity of the solutions in the various methods .
the existence of the spectral representation is connected to the branch cut of physical amplitudes on the negative coupling ( @xmath3 ) axis .
this branch cut is due to the non - existence of the theory at negative couplings : the path integral diverges due to a potential unbounded from below . in a higher dimensional field theory
this has a simple physical interpretation : the vacuum is unstable and , after tunneling through a barrier , the system rolls down the potential @xcite . for weak coupling
the semi - classical approximation is valid and the tunneling is exponentially suppressed , giving an imaginary contribution @xmath4 to the vacuum persistence amplitude which is inherited by the green function .
this exponential behaviour can be seen in the spectral function we obtain .
dyson @xcite argued that a very similar phenomenon occurs in quantum electrodynamics ( qed ) .
we briefly reiterate this argument . in qed physical observables
are calculated in a perturbation series of the form @xmath5 where @xmath6 is the charge of the electron .
now if one imagines a world where @xmath7 , i.e. like charges attract , it is easy to see that the ordinary vacuum is unstable to the production of many electron - positron pairs which separate into clouds of like - charged particles . at weak coupling
there is a large tunneling barrier to overcome because one must pay for the rest mass of the pairs and separate them far enough for the wrong - sign coulomb potential to compensate .
thus there is a finite but exponentially suppressed rate of vacuum decay .
a taylor series expansion in @xmath8 can not capture this non - analyticity so the perturbation series must be divergent .
similarly , the perturbation series in @xmath3 diverges for the toy model considered here .
pad approximants are more effective because they can develop isolated poles in the complex @xmath3 plane , however they struggle to capture the strong coupling behaviour at any fixed order in the approximation .
pad approximants are better able to capture the non - analyticities of the borel transform , however , and the widely used combination borel - pad approximants give a better global approximation .
this occurs because the pad approximated bore transform has poles in the borel plane , which lead to branch cuts when the laplace transform is taken to return to physical variables .
similarly , the self - consistent 2pi approximations develop branch point non - analyticities and approximate the exact green function rather well in the entire complex @xmath3 plane already at the leading non - trivial truncation . however , the branch cuts in the 2pi case arise because the 2pi green function obeys self - consistent equations of motion , and is connected to the existence of unphysical solution branches .
a question that naturally arises is : how do these methods compare ?
both 2pi and borel - pad methods have the ability to accurately represent non - analyticities of the exact theory and so out - perform other methods .
however , we conjecture that self - consistently derived equations of motion `` know more '' about the analytic structure of the underlying theory than do the generic borel - pad approximants
. hence we test the hypothesis that the 2pi methods should be more accurate that borel - pad and , indeed , find this to be the case , at least in certain regimes .
the theory discussed here , although admittedly a toy model , also has physical relevance .
independent from us , beneke and moch found this toy model as the theory governing the zero mode of scalar fields in euclidean de sitter space @xcite .
they performed an analysis very similar to ours , finding that a non - perturbative treatment is necessary and comparing 2pi and ( borel-)pad resummed approximations .
however , they present this analysis very briefly as part of a larger discussion of scalar fields in de sitter space .
further , their comparison of the 2pi and resummed techniques , while correct as far as we can tell , is not very detailed . here
we present detailed discussion of the interplay between 2pi effective actions and various resummation techniques .
our use of the spectral function to quantify the non - analyticities present in the green function in aid of this comparison is , as far as we are aware , a new aspect .
this paper has a somewhat pedagogical flavour , and readers familiar with field theory , borel summation and pad approximants can skim through sections [ sec : exact - solution - of ] and [ sec : perturbation - theory - and - resummations ] where these topics are discussed , pausing only to pick up our notation and our derivation of the spectral function for the exact theory ( section [ sec : exact - solution - of ] ) and the pad resummed theory ( section [ sub : pade - approximation ] ) . in section [ sec:2pi - approximations ] we compute the 2pi effective action , green function and corresponding spectral function for the theory . in section [ sec : hybrid-2pi - pade ] we introduce , for the first time to our knowledge , a hybrid 2pi - pad scheme and show that it accelerates convergence at weak coupling while having the correct behaviour at strong coupling , like 2pi approximations but not the usual pad method .
finally in section [ sec : discussion ] we draw our conclusions .
we consider a euclidean qft in zero dimensions , i.e. a probability theory for a single real variable @xmath9 given by the partition function @xmath10=n\int_{-\infty}^{\infty}\mathrm{d}q\exp\left(-\frac{1}{2}m^{2}q^{2}-\frac{1}{4!}\lambda q^{4}-\frac{1}{2}kq^{2}\right)\label{eq : partition - function}\ ] ] in the presence of a source @xmath11 for the two point function .
@xmath12 is a normalizing factor chosen so that @xmath13=1 $ ] .
this theory has been discussed before in the context of exact and non - perturbative methods in field theory ( see , e.g. , @xcite and references therein ) because , despite the absence of spacetime , the theory possesses a perturbative expansion in terms of feynman diagrams with the same combinatorial structure as more realistic theories .
it was also discussed in @xcite as an effective field theory for scalar field zero modes in euclidean de sitter space .
we restrict attention to the @xmath14 theory since , though the @xmath15 theory exists and may be interesting for other purposes , it possesses no sensible weak coupling limit and in zero dimensions does not give a broken symmetry phase anyway.$ ] only really depends on the ratio @xmath16 , so the only sensible definition of weak coupling is @xmath17 , a limit which sends the entire support of the integral to @xmath18 if @xmath15.]^,^ in the absence of symmetry breaking we may omit a source term for @xmath9 . the integral diverges for @xmath19 , but we can define the value of @xmath20 $ ] at complex @xmath3 by analytic continuation . then @xmath20 $ ] possesses a branch cut along the negative @xmath3 axis .
physically , this signals the instability of the negative @xmath3 vacuum due to tunneling away from the local minimum at @xmath21 to @xmath22 followed by rolling down the inverted quartic potential which is unbounded from below .
the branch point at @xmath23 means that the weak coupling perturbation series has zero radius of convergence .
this is similar to the behaviour which exists in most theories of physical interest as argued by dyson @xcite .
the analytic behaviour of the integrand for @xmath24 is shown in figure [ fig : z - integrand ] .
one can see that the integrand has a maximum at @xmath25 and saddle points at @xmath26 .
@xmath20 $ ] for the @xmath14 theory can be obtained from integration along the real axis , while integration along the imaginary axis gives the @xmath15 theory .
similarly , taking the reciprocal of the integrand sends @xmath27 , @xmath28 and also reverses the colour map in figure [ fig : z - integrand ] ( changing maxima to minima and vice versa ) . from this
we can see that , as expected , the partition function diverges for @xmath29 and theory makes no sense .
we introduce the conveniently rescaled variables @xmath30 and @xmath31 and obtain @xmath10=\frac{2n}{m}\sqrt{\rho\left(1+k\right)}\exp\left(\rho\left(1+k\right)^{2}\right)\mathrm{k}_{1/4}\left(\rho\left(1+k\right)^{2}\right),\ ] ] where @xmath32 is a modified bessel function of the second kind .
this expression is valid so long as @xmath33 which extends the definition to the entirety of the cut @xmath3-plane .
the normalization factor is @xmath34 and finally @xmath10=\sqrt{1+k}\exp\left(\rho\left[\left(1+k\right)^{2}-1\right]\right)\frac{\mathrm{k}_{1/4}\left(\rho\left(1+k\right)^{2}\right)}{\mathrm{k}_{1/4}\left(\rho\right)}.\ ] ] @xmath20 $ ] is used to compute the expectation value of physical observables @xmath35 , by the standard trick of differentiating under the integral and then removing the source : @xmath36\right|_{k=0}.\end{aligned}\ ] ] we also define the generating function @xmath37=-\ln z\left[k\right]\label{eq : w - generating - function}\ ] ] and note that averages are found by taking derivatives of @xmath38 .
for example , @xmath39 & = \frac{1}{2}\left\langle q^{2}\right\rangle _ { k}\equiv\frac{1}{2}\bar{g},\label{eq : g - def}\\ \frac{\partial^{2}}{\partial k^{2}}w\left[k\right ] & = -\frac{1}{4}\left(\left\langle q^{4}\right\rangle -\left\langle q^{2}\right\rangle ^{2}\right)_{k}\equiv\frac{1}{4}\left(v^{\left(4\right)}\bar{g}^{4}-2\bar{g}^{2}\right),\label{eq:4pt - vertex - def}\end{aligned}\ ] ] where the subscript @xmath11 indicates the average is taken at a fixed value of @xmath11 .
( note that @xmath40 $ ] is not the _ connected _ generating function as usually defined because @xmath11 is a _
two_-point source . )
@xmath41 and @xmath42 are the proper two and four point functions respectively .
to lowest order in perturbation theory and with @xmath43 , @xmath44 and @xmath45 .
the exact value of @xmath40 $ ] is easily obtained directly from the definition , giving @xmath46 & = -\frac{1}{2}\ln\left(1+k\right)-\rho\left[\left(1+k\right)^{2}-1\right]-\ln\mathrm{k}_{1/4}\left(\rho\left(1+k\right)^{2}\right)+\ln\mathrm{k}_{1/4}\left(\rho\right).\end{aligned}\ ] ] by direct differentiation we obtain the exact two point correlation function @xmath47,\end{aligned}\ ] ] or for the original ( @xmath43 ) theory , @xmath48 like @xmath49 , @xmath41 possesses a branch cut discontinuity from @xmath23 to @xmath50 . at @xmath23
one obtains the usual free ( gaussian ) theory result @xmath51 . in the strong coupling limit , @xmath52 , @xmath53 \lambda^{-1/2 } + \mathcal{o}\left(\lambda^{-1}\right)$ ] .
@xmath41 is shown in figure [ fig : gexact ] , from which the branch cut is obvious .
this can also be seen in more detail in the complex @xmath3 plane as shown in figure [ fig : gexact - analytic - behaviour ] .
one sees that not only does @xmath41 possess a branch cut , it is analytic in the cut plane and is in fact a herglotz - nevanlinna function ( i.e. @xmath54 where @xmath55 is complex conjugation ) .
this means that @xmath41 has a nice integral representation which we derive now in order to quantify the branch cut .
@xmath41 from as a function of @xmath3 for @xmath56 .
the sign of the imaginary part depends on whether one approaches the @xmath29 cut from above or below .
note the imaginary part is exponentially suppressed near the origin because the vacuum decay process is non - perturbative . ]
we obtain the integral representation for @xmath41 using the cauchy formula @xmath57 where the contour @xmath58 circles @xmath3 in the counter - clockwise direction and avoids the cut .
deforming the contour to run on the circle at infinity and around the cut and using @xmath59 as @xmath60 , we can write the integral in terms of a spectral function @xmath61\theta\left(s\right)/2\pi i=\mathrm{im}\left[\bar{g}\left(-s - i\epsilon\right)\right]\theta\left(s\right)/\pi$ ] where @xmath62 is the heaviside step function , such that @xmath63 we find @xmath64}\theta\left(\lambda\right),\label{eq : spectral - function - exact}\ ] ] where @xmath65 are modified bessel functions of the first kind . a purely formal device that gives information about the analytic structure of @xmath41 .
] @xmath66 is shown in figure [ fig : sigma - exact ] .
spectral function @xmath66 of for @xmath56 . ]
at small coupling one often uses perturbation theory in @xmath3 which proceeds by expanding the exponential @xmath67 & = n\int_{-\infty}^{\infty}\mathrm{d}q\exp\left(-\frac{1}{2}\left(m^{2}+k\right)q^{2}-\frac{1}{4!}\lambda q^{4}\right)\nonumber \\ & = n\int_{-\infty}^{\infty}\mathrm{d}q\sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda q^{4}\right)^{n}\exp\left(-\frac{1}{2}\left(m^{2}+k\right)q^{2}\right)\nonumber \\ & \sim\sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda\right)^{n}n\int_{-\infty}^{\infty}\mathrm{d}qq^{4n}\exp\left(-\frac{1}{2}\left(m^{2}+k\right)q^{2}\right)\nonumber \\ & = \sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda\right)^{n}n2^{2n+\frac{1}{2}}\gamma\left(2n+\frac{1}{2}\right)\frac{1}{\left(m^{2}+k\right)^{2n+\frac{1}{2}}}.\end{aligned}\ ] ] in the third line we have formally interchanged the sum and integral , leading to an asymptotic rather than convergent series for @xmath20 $ ] .
again , @xmath12 is determined by @xmath13=1 $ ] , giving @xmath68^{-1},\ ] ] so @xmath10=\frac{\sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda\right)^{n}2^{2n}\gamma\left(2n+\frac{1}{2}\right)\frac{1}{\left(m^{2}+k\right)^{2n+\frac{1}{2}}}}{\sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda\right)^{n}2^{2n}\gamma\left(2n+\frac{1}{2}\right)\frac{1}{\left(m^{2}\right)^{2n+\frac{1}{2}}}},\ ] ] and @xmath37=-\ln\sum_{n=0}^{\infty}\frac{1}{n!}\left(-\frac{1}{4!}\lambda\right)^{n}2^{2n}\gamma\left(2n+\frac{1}{2}\right)\frac{1}{\left(m^{2}+k\right)^{2n+\frac{1}{2}}}+\text{const}.\ ] ] from this we find @xmath69 the perturbative approximations @xmath70 are the @xmath71 truncations of this series .
the first few are shown compared to the exact @xmath41 in figure [ fig : g - pert - series ] .
note that these are simply the low order taylor series approximations to @xmath41 .
these approximations apparently converge poorly to the exact solution , and in fact we will shortly show that the series diverges . the series for @xmath41 can be described in terms of feynman diagrams by the following rules : 1 .
draw all connected graphs with two external lines ( i.e. lines with one end not connected to any vertex ) constructed from lines and four point vertices .
2 . associate to each line a factor @xmath72 .
3 . associate to each vertex a factor @xmath73 .
4 . divide by an overall symmetry factor being the order of the symmetry group of the diagram under permutations of lines and vertices .
we illustrate these rules by giving the first few terms in @xmath41 : perturbation - theory - propagator @xmath74 @xmath40 $ ] can be written as a similar diagrammatic series in terms of connected vacuum diagrams ( i.e. , those with no external lines ) .
perturbative approximations to @xmath41 up to @xmath75 for @xmath56 , compared to the exact solution . ]
the series has the form @xmath76 where the coefficients asymptotically obey @xmath77 as @xmath78 , thus the radius of convergence of the series is zero .
this is consistent with the fact that we are perturbing around a branch point of the exact solution : no approximation of @xmath41 in terms of analytic functions can converge at @xmath23 because @xmath20 $ ] is itself undefined for @xmath19 .
the terms of the series start to increase when @xmath79 , i.e. @xmath80 , meaning the series is useful for @xmath17 but fails immediately for a moderately strong coupling @xmath81 .
this is typical asymptotic series behaviour as shown in figure [ fig : gpert - asymptotic - series ] .
extrapolating perturbation theory to strong coupling @xmath82 is simply impossible , although the exact solution is well behaved there .
( in fact @xmath41 can be expanded as @xmath83 , displaying explicitly the branch point at @xmath84 . )
relative error @xmath85of the @xmath86-th perturbative approximation to @xmath41 for @xmath56 and @xmath87 , showing the decreasing then increasing behaviour typical of asymptotic series . ]
we have seen that the series expansion for @xmath41 diverges for all @xmath88 .
this is typical of perturbation series and usually signals some singularity of the exact solution for unphysical values of @xmath3 . in our case , indeed , the theory does not exist for @xmath19 and the exact solution possesses a branch cut on the negative @xmath3 axis , a feature which can not be reproduced in any order of perturbation theory .
however , the perturbation series is asymptotic and does contain true information about the exact solution , even if @xmath3 is large enough that the series is not useful practically .
because of the ubiquity of this phenomenon , mathematicians have invented a number of series summation techniques which assign a finite value to certain types of divergent series and which obey certain consistency properties ( e.g. the value assigned to a convergent series is just its nave sum ) .
here we investigate borel summation , which is capable of summing factorially divergent series like .
suppose that @xmath89 is a divergent series but that the _ borel transform _ of the series , defined as @xmath90 converges for sufficiently small @xmath91 .
then , if the integral @xmath92 exists the _ borel sum _ @xcite of the divergent series
is defined as @xmath93\equiv b\left(1\right)$ ] .
this definition is justified by substituting the series for @xmath94 into the integral and evaluating term - wise and noting that @xmath95 .
the main drawback of borel summation is that one must know the precise form of @xmath96 for all @xmath86 to compute @xmath97 , which is rarely the case in field theory .
for this reason borel summation can not be usefully applied directly .
however , one may use pad approximants as discussed in the next section to recast the borel transform in a useful way .
we note that the key to borel - summability of the perturbation series is the alternating sign @xmath98 of the @xmath86-th order term . to see this consider the two series @xmath99 which differ only by the alternating sign .
the borel transforms @xmath100 are @xmath101 and the borel sums are @xmath102=b_{1}\left(1\right ) & = \int_{0}^{\infty}\frac{\mathrm{e}^{-t}}{1+\lambda t}\mathrm{d}t,\\
\mathcal{b}\left[s_{2}\right]=b_{2}\left(1\right ) & = \int_{0}^{\infty}\frac{\mathrm{e}^{-t}}{1-\lambda t}\mathrm{d}t.\end{aligned}\ ] ] in the first case the integral exists and @xmath103=\lambda^{-1}\mathrm{e}^{1/\lambda}\gamma\left(0,\frac{1}{\lambda}\right)$ ] where @xmath104 is the incomplete gamma function .
however , the second integral hits a pole at @xmath105 .
there is no natural prescription for avoiding the pole , which leads to an ambiguity in the sum of @xmath106 .
this is a non - perturbative ambiguity called a _
renormalon _ @xcite . in every known case where this arises in field theory the renormalon
is connected to a non - perturbative finite action solution of the field equations , i.e. an instanton or soliton , and a correct evaluation of the path integral which sums over _ all _ saddle points ( not just perturbative ones ) removes the ambiguity .
key to the practical application of borel summation is the location and classification of all renormalons in a given theory @xcite .
sophisticated techniques have been developed to deal with this situation which are beyond the scope of this paper @xcite .
borel summation on its own has limited usefulness in practice because one often only knows a few low order terms of perturbation theory , and the potential existence of renormalon singularities .
there exists another technique which often improves perturbation series and is far more useful in practice ( and is often combined with borel summation ) .
pad approximation approximates a function by rational polynomials which generally converge rapidly , are very useful for numerical computation and help to estimate the location of singularities of the function in the complex plane .
many software packages include standard routines for evaluating pad approximants , for instance the padeapproximant function in mathematica . in this section
we apply pad approximants directly to the green function and then to the borel transform .
we find that the latter approach is clearly the better one .
the @xmath107-pad approximant of a function @xmath108 is given by @xcite @xmath109 where without loss of generality one takes @xmath110 .
the remaining @xmath111 coefficients are chosen so that the taylor series of @xmath112 matches the perturbation series up to @xmath113 . due to the denominator ,
pad approximants develop poles in the complex @xmath91-plane , allowing the close approximation of more singular functions than taylor series are capable of .
examples of the use of pad approximants in field theory can be found in @xcite and references therein .
now we find the pad approximants to @xmath114 with @xmath115 .
here we restrict attention to the approximants where @xmath116 .
this guarantees that @xmath117 as @xmath52 .
if we had used the usual diagonal approximants ( @xmath118 ) we would find an unphysical constant term @xmath119 as @xmath52 .
note that it is impossible to match the true @xmath120 behaviour of @xmath41 as @xmath52 using pad approximants centred on the origin .
the best that is possible in this limit is @xmath121 .
( as it happens , using @xmath122 does not allow one to resolve this issue : the same approximants are found only with @xmath123 everywhere in place of @xmath91 .
this is because the @xmath120 behaviour is due to the branch point at infinity , which is infinitely far from the origin where the pad approximants are matched to perturbation theory .
low order pad approximants can extract information about the branch point near the origin , but evidently not the one at infinity . ) the first five approximants for @xmath114 are shown in table [ tab : pade - approximants ] and the first three are plotted in figure [ fig : pade - approx ] with comparison to the exact @xmath41 .
note that the existence of the integral representation for @xmath41 implies that @xmath41 is a _
stieltjes _ function , meaning one can prove convergence properties for the pad approximants as @xmath124 , though we are not concerned with this analysis here ( see @xcite for details ) .
.[tab : pade - approximants]first few pad approximants for @xmath114 . [ cols="<,<",options="header " , ] first three pad approximants to @xmath41 .
note the simple poles developed for @xmath29 . ]
the @xmath125-pad approximant can also be written as @xmath126 where @xmath127 and @xmath128 are the @xmath129-th residue and pole respectively .
note that since all of the coefficients in the denominators of table [ tab : pade - approximants ] are positive and real , all of the poles must either be on the negative real axis or they must be complex and come in complex conjugate pairs .
numerical experiments suggest that all the poles lie on the negative @xmath3 axis , though we do not know a proof of this for all @xmath12 . assuming this is generally true
, pad approximants give a representation of @xmath41 which approximates the continuous spectral function by a sum of delta functions @xmath130 as @xmath131 the poles become denser and fill the negative @xmath3 axis , eventually merging into a continuous branch cut .
similarly the spectral function turns into a dense sum of delta functions which , when considered acting on any sufficiently smooth test function , smooths into a continuous function .
the first few @xmath132 are shown next to the exact spectral function in figure [ fig : pade - spectral - function ] for comparison .
pad approximations to the spectral function @xmath66 consist of an increasingly dense set of delta functions .
( note the delta functions have been smoothed for visual purposes only . ) ] now we consider pad approximation of the borel transform of @xmath41 .
first we note the following connections between the borel transform @xmath133 and @xmath41 and @xmath134 : @xmath135{\mathcal{l}^{-1}}\phi\left(x\right)\xrightarrow[x\to s]{\mathcal{l}}s^{-1}\bar{g}\left(\frac{\lambda}{s}\right).\ ] ] that is , the green function is related to the laplace transform of the borel transform , while the spectral function is related to the inverse laplace transform of the borel transform .
these relations can be shown using the definitions of @xmath133 and @xmath134 and the integral representation of the ( inverse ) laplace transform .
this allows us to extract the spectral function directly from the borel transform .
note that each pole of @xmath133 yields by the inverse laplace transform a term of the form @xmath136 in @xmath134 , where @xmath137 is controlled by the location of the pole .
the general borel - pad approximation for @xmath134 is a superposition of terms of this form .
we show the first few approximants to @xmath41 and @xmath134 in figures [ fig : gbar - padeborel ] and [ fig : sigma - padeborel ] respectively .
one sees that the low order borel - pad green functions are reasonably accurate ( within a few percent for @xmath138 ) but the spectral functions are not approximated particularly well .
certain approximants to @xmath134 oscillate erratically and even become negative for certain values of @xmath3 . except for the fact that the best approximant is the highest order
one plotted , there is no clear sense in which the borel - pad approximations appear to converge to @xmath134 .
however , even this bad approximation is at least a continuous function , as opposed to a sum of delta functions .
borel - pad approximations to the green function @xmath139 ( ratio of approximant / exact ) .
approximants are calculated by numerical integration of . ]
borel - pad approximations to the spectral function @xmath66 ( ratio of approximant / exact ) .
note that the ( 2,2 ) approximant is off scale . ]
the 2pi effective action is a functional of the green function @xmath140 which is defined by the legendre transform @xcite @xmath141=w\left[k\right]-k\partial_{k}w\left[k\right]\ ] ] where @xmath11 is solved for in terms of @xmath142 .
@xmath143 $ ] obeys the equation of motion @xmath144=-\frac{1}{2}k.\ ] ] the standard derivation of the 2pi action @xcite gives in this case @xmath141=\frac{1}{2}\ln\left(g^{-1}\right)+\frac{1}{2}m^{2}g+\gamma_{2\mathrm{pi}}+\mathrm{const.},\ ] ] where @xmath145 is ( minus due to euclidean conventions ) the sum of two particle irreducible vacuum graphs , i.e. those graphs which do not fall apart when any two lines are cut , where the lines are given by @xmath142 and vertices by @xmath73 . explicitly : twopi - diagrams @xmath146{15mm}{\begin{fmfgraph}(15,20 ) \fmfleft{l}\fmfright{r } \fmf{plain , left}{l , v , r , v , l } \fmfdot{v } \end{fmfgraph } } \ + \frac{1}{4!2}\
\parbox{10mm}{\begin{fmfgraph}(10,20 ) \fmfleft{l}\fmfright{r } \fmf{plain , left}{l , r , l } \fmf{plain , left=0.5}{l , r , l } \fmfdot{l , r } \end{fmfgraph } } \ + \frac{1}{2 ^ 3 3!}\parbox{10mm}{\begin{fmfgraph}(10,10 ) \fmfsurroundn{v}{3 } \fmfdotn{v}{3 } \fmfcyclen{plain , right=0.25}{v}{3 } \fmfrcyclen{plain , right=0.25}{v}{3 } \end{fmfgraph } } + \mathcal{o}\left(\lambda^4\right).\end{aligned}\ ] ] the equation of motion in the absence of sources is @xmath147 which has the diagram expansion 2pi - eom @xmath148 notice that there is a dramatic reduction in the number of graphs of a given order compared to perturbation theory due to the two - particle irreducibility .
this is one of the major benefits of 2pi approximation schemes in realistic theories , such as gauge theories , where the feynman diagrams proliferate rapidly .
the equation of motion is dyson s equation and the second term on the right hand side , @xmath149 , represents the exact one - particle irreducible self - energy of the propagator @xmath142 .
this can be put into the usual form of a dyson equation by noting @xmath150 and multiplying both sides by @xmath151 : @xmath152 iterating this gives the infinite series @xmath153 the main difference between this and the usual perturbative dyson equation is that @xmath154 contains exact propagators @xmath142 rather than @xmath155 . inserting the expression above for @xmath142 into @xmath154 one finds that , even if one retains only a finite number of 2pi diagrams in @xmath156 , @xmath154 contains an infinite series of perturbative self - energy graphs .
this is the motivation in the literature for talking about 2pi as a resummation method
. by power counting ( or counting line ends in the corresponding diagrams ) we find @xmath157 .
it is possible to derive the @xmath158 by considering the symmetry factors of the two particle irreducible feynman diagrams , but we can also obtain the @xmath158 in a simple automated way using knowledge of the exact solution @xmath159 . substituting @xmath41 into the equation of motion and the expansion for @xmath156 , expanding about @xmath23 and matching powers of @xmath3
we can determine @xmath160 we do not know of any closed form expression for either the coefficients of this series or its sum ( implicit analytic expressions can be derived ; however , these require the inversion of @xmath161 for @xmath162 , which is not known to us in closed form ) .
however , after the first few terms the coefficients seem to be well approximated by @xmath163 , the same as for the perturbation series .
this has the hallmark of an asymptotic series .
like perturbation theory , the 2pi series does not converge .
the first non - trivial contribution to the equation of motion gives @xmath164 where the subscript @xmath165 indicates terms of order @xmath166 have been kept .
this has two solutions @xmath167 one of these solutions is unphysical and we must choose the @xmath168 sign . as @xmath169 , @xmath170 which matches perturbation theory up to @xmath171 terms as expected .
however , unlike perturbation theory , the strong coupling limit @xmath52 exists and gives @xmath172 this series has the correct form in powers of @xmath173 , though the leading coefficient is incorrect by @xmath174 , and the sub - leading coefficients are incorrect by @xmath175 , @xmath176 and @xmath177 etc .
nevertheless , it is remarkable to achieve any accuracy at all given the simple nature of the approximation and the fact that @xmath156 was truncated at leading order in @xmath3 !
@xmath178 is a much more uniform approximation to @xmath41 than the perturbative approximation @xmath179 .
this result is possible because of the branch cut @xmath178 possesses on the negative @xmath3 axis .
the discontinuity across the cut @xmath180 gives the spectral function @xmath181 this is a far better approximation to the exact spectral function than obtained from any of the other techniques , as can be seen from figure [ fig : sigma-2pi-2loop ] .
comparison of exact spectral function @xmath66 with the two loop 2pi approximation @xmath182 for @xmath56 .
already , the simplest nontrivial 2pi truncation gives a much better approximation than perturbation theory or pad approximants to arbitrary order . ]
we show all of the first order approximations in figure [ fig : gbar - all - first - order - approxs ] .
note that for @xmath183 the best approximation is the 2pi , followed by the borel - pad , then by pad , then perturbation theory last of all .
the situation for negative @xmath3 is complicated .
the best approximation overall is the 2pi , though it has an unphysical cusp where the branch cut starts ( @xmath184 in figure [ fig : gbar - all - first - order - approxs ] ) .
it appears the 2pi approximation trades sensitivity to the exponentially small portion of @xmath134 in exchange for a better global approximation .
the pad approximation is good at small negative @xmath3 but hits a pole at @xmath185 and there loses all validity .
the borel - pad approximation is smooth and more accurate than the 2pi near the peak of @xmath41 , but eventually becomes invalid , even negative , at sufficiently large negative @xmath3 ( @xmath186 ) .
comparison of ( the real part of ) the exact green function @xmath41 to each of the approximations discussed to first non - trivial order in each case . ] at @xmath86-th order , is a degree @xmath187 polynomial in @xmath142 which has @xmath187 roots , only one of which is physical . for @xmath188
there are analytical expressions for the roots , though they are very bulky , and for @xmath189 must be solved numerically . picking out the correct root for a given value of @xmath3 is tricky in general and we leave this exercise to the reader .
however , one can see on general grounds that truncations of give a singular perturbation theory in @xmath3 : at @xmath23 the physical solution starts at @xmath155 and the spurious solutions flow in from infinity as inverse powers of @xmath3 as @xmath3 increases to finite values . at strong coupling
the roots generically approach each other and one must carefully track them through the complex plane .
we suspect that a resummation of the 2pi series would remove some or all of these spurious solutions , though we do not have a proof of this .
we examine two potential methods to achieve this in the next section and find mixed results .
since the series of 2pi diagrams is asymptotic , we may consider using a series summation method to improve convergence .
note that this is logically independent of the resummations embodied in the 2pi approximation itself .
we may perform pad summation of the action term @xmath156 or the equation of motion term @xmath190 .
we consider both .
first , consider pad summation of the action which matches the series expansion up to order @xmath191 : @xmath141=\frac{1}{2}\ln\left(g^{-1}\right)+\frac{1}{2}m^{2}g+\frac{\sum_{n=0}^{n}a_{n}^{\left(\gamma\right)}\left(\lambda g^{2}\right)^{n}}{\sum_{n=0}^{m}b_{n}^{\left(\gamma\right)}\left(\lambda g^{2}\right)^{n}}+\mathrm{const.}\ ] ] the equation of motion becomes @xmath192^{2}}.\label{eq : hybrid - attempt-1}\end{aligned}\ ] ] multiplying by the denominator this becomes a polynomial equation of degree @xmath193 as expected .
this equation will have @xmath194 spurious solutions as does the ordinary 2pi approximation of matching order . to take a specific example ,
consider @xmath156 up to @xmath195 and the matching @xmath196- and @xmath197-pad approximants : @xmath198 the solution resulting from any of these versions of @xmath156 can not be written in closed form , however numerical solutions are shown in figure [ fig : hybrid - solution ] for moderately small @xmath3 ( for @xmath3 outside of this range , different roots become relevant ) .
we see that the hybrid solutions are more accurate than the fourth order 2pi solution , although the level of improvement depends strongly on the type of pad approximant employed .
the trade - off is a loss of accuracy at large @xmath3 , as shown in figure [ fig : hybrid - solution - large - lambda ] for the best roots we can find .
each approximation decays with the correct leading @xmath173 power , however they differ by @xmath199 factors which are not negligible .
remarkably , the best approximation of those shown is the @xmath200 2pi approximation . for the greater computational expense
the @xmath196-hybrid approximation is not noticeably an improvement at large coupling .
the apparent clustering of the @xmath195 2pi and @xmath197-hybrid approximations away from the exact value as @xmath52 probably stems from the minus sign of the leading term in @xmath190 in these approximations , suggesting that one should not use approximations with this property .
comparison of exact @xmath41 , first and fourth order 2pi and@xmath196- and @xmath197 - 2pi - pad hybrid solutions at small coupling .
the @xmath196-hybrid solution lies almost on top of the exact solution . ]
comparison of exact @xmath41 , first and fourth order 2pi and@xmath196- and @xmath197 - 2pi - pad hybrid solutions at large coupling . ]
now we try pad summing the equation of motion : @xmath201 where the explicit factor of @xmath202 on the right hand side simply ensures the self - energy is an odd function of @xmath142 as it must be .
this equation of motion is of degree @xmath203 , which for typical choices of @xmath12 and @xmath204 results in a rough halving of the number of spurious solutions . to obtain an analytic result
we consider the first non - trivial approximant , i.e. @xmath205 . the resulting equation of motion is @xmath206 which agrees with the usual 2pi equation of motion up to terms of order @xmath207 , however it only has two spurious solutions instead of three .
the physical solution obtained by matching at small @xmath3 is @xmath208 where @xmath209 indeed , this matches perturbation theory up to @xmath207 terms .
however , the large @xmath3 behaviour in this approximation is pathological : @xmath210 as @xmath52 .
this reflects the existence of an unphysical branch cut on the positive @xmath3 axis starting at @xmath211 , which @xmath212 has in addition to the expected cuts on the negative axis .
the existence of this cut also renders the derivation of the spectral function invalid , meaning that @xmath212 can not be written in the form .
it is not clear at this stage whether other forms of pad summed equations of motion lack these pathologies .
we leave further investigation of hybrid approximation schemes to future work .
two - particle irreducible effective actions are the subject of a rich literature and have found applications in diverse areas from early universe cosmology to nano - electronics ( e.g. @xcite ) .
the virtues of approximation schemes built on 2pi effective actions are often explained in terms of a resummation of an infinite series of perturbative feynman diagrams which are encapsulated in the non - perturbative green function @xmath142 , from which the 2pi diagram series is built .
however , the 2pi effective action is not a resummation scheme : it is a self - consistent variational principle .
the definition of the 2pi effective action in terms of the legendre transform is crucial for the self - consistency of the scheme .
the immediate practical consequence is that any modification of the 2pi effective action which does not derive from a consistent modified variational principle is very likely to be inconsistent .
so for instance , the consistency of recent attempts to improve the symmetry properties of 2pi effective actions @xcite is guaranteed by the existence of a suitable constrained variational principle , however , ad hoc attempts to modify the equations of motion to satisfy symmetries will fail . in this work
we have pointed out the distinction between resummation and self - consistent approximations using an exactly solvable zero dimensional `` field '' theory .
the perturbation theory has zero radius of convergence due to a branch cut on the negative coupling ( @xmath3 ) axis , a fact which is invisible to perturbation theory .
the theory is borel summable , with the borel sum giving the exact answer .
however , borel summation alone is not usually this helpful in practice .
pad approximants well describe the green function at weak coupling , though not at strong coupling . however , the combination of borel and pad approximation yields an effective global approximation scheme for the green function .
the two point green function of this theory admits a nice integral representation in terms of the spectral function ( i.e. discontinuity of the branch cut ) which allows us to see that the pad approximants improve perturbation theory by allowing the spectral function to be approximated as a sum of delta functions and the borel - pad method gives a continuous , albeit erratic and inaccurate , approximation to the spectral function .
the 2pi approximation scheme surpasses perturbation theory , pad and borel - pad approximants already at the leading non - trivial truncation . like the borel - pad method
, 2pi approximations can develop branch points and represent the spectral function by a continuous distribution .
however , the 2pi approximation is quantitatively superior at the leading truncation .
we speculate that this is because the borel - pad method is a widely applicable general `` black box '' method , however the self - consistent 2pi equations of motion are derived within a _ particular _ field theory of interest .
this gives the 2pi method `` insider information , '' from which it should be able to construct a better approximation .
this comes at the cost of spurious solutions which must be eliminated and the added difficulty of finding the 2pi effective action in the first place .
finally we introduced ( for the first time , to our knowledge ) a hybrid 2pi - pad scheme , using pad approximants to partially resum the 2pi diagram series .
the quality of the result depends strongly on the type of pad approximant used , with the best result in our case for the diagonal approximant .
this hybrid approximation performs considerably better than the comparable 2pi approximation at weak coupling , though not noticeably better at strong coupling . during proofing we became aware of a work by kleinert @xcite in some ways similar to our own .
in @xcite , kleinert generalizes and reformulates the feynman - kleinert _ variational perturbation theory _ and compares it to the 2pi effective action ( called by the non - standard name `` bilocal legendre transform '' ) for the toy model discussed here .
he finds that the variational perturbation theory out - performs the 2pi method and notes especially the failure of the 2pi series to converge uniformly for all couplings ( this is not a surprise in light of the fact that the 2pi series is asymptotic ) .
our work differs in scope from his in two major ways .
first , we compare the 2pi method to traditional resummation methods and not to other variational methods .
second , kleinert focuses on the value of the effective action itself evaluated at its extremum , while we focus on the correlation function . in particular ,
we have focused on the analytic structure of the correlation function in the complex coupling plane , and introduced the spectral representation to aid this discussion .
a study of the behaviour of these quantities in variational perturbation theory is an intriguing prospect for future work .
( an intermediate result of kleinert s work is directly applicable to our discussion around : equation 110 of @xcite gives a non - linear recurrence relation for the coefficients of the 2pi self - energy . )
there are several other natural directions for extension of this work .
one would be the calculation and comparison of higher orders in the 2pi and ( borel-)pad approximations , although it is doubtful what new insights might come from this .
it would be straightforward to extend this work to consider 4pi effective actions , which depend on a self - consistent vertex function @xmath42 in addition to @xmath142 . in and performing the legendre
transform with respect to both @xmath213 and @xmath11 .
] however , the extension to higher order @xmath86pi effective actions requires the introduction of new terms in the exponent of which makes the problem no longer exactly solvable .
a natural direction to pursue is borel - pad summation of the 2pi generating functional itself .
this opens the possibility of eliminating spurious solutions to the 2pi equations of motion and enhancing the sensitivity of the 2pi method to the exponentially small regions of the spectral function , hence eliminating unphysical cusps .
we are planning an investigation of these themes for the quantum anharmonic oscillator as a stepping stone to more physically interesting field theories .
t. d. lee , c. n. yang , http://link.aps.org/doi/10.1103/physrev.117.22[many-body problem in quantum statistical mechanics .
iv . formulation in terms of average occupation number in momentum space ] , physical review 117 ( 1 ) ( 1960 ) 2236 . http://dx.doi.org/10.1103/physrev.117.22 [ ] .
j. m. luttinger , j. c. ward , http://link.aps.org/doi/10.1103/physrev.118.1417[ground-state energy of a many - fermion system .
ii ] , physical review 118 ( 5 ) ( 1960 ) 14171427 .
http://dx.doi.org/10.1103/physrev.118.1417 [ ] .
g. baym , http://link.aps.org/doi/10.1103/physrev.127.1391[self-consistent approximations in many - body systems ] , phys . rev . 127 ( 1962 ) 13911401 .
[ ] . http://link.aps.org/doi/10.1103/physrev.127.1391 j. cornwall , r. jackiw , e. tomboulis , http://link.aps.org/doi/10.1103/physrevd.10.2428[effective action for composite operators ] , phys .
d 10 ( 8) ( 1974 ) 24282445 .
http://dx.doi.org/10.1103/physrevd.10.2428 [ ] .
j. berges , http://scitation.aip.org/content/aip/proceeding/aipcp/10.1063/1.1843591[introduction to nonequilibrium quantum field theory ] , aip conf .
739 ( 1 ) ( 2004 ) 362 . http://arxiv.org/abs/hep-ph/0409233 [ ] , http://dx.doi.org/http://dx.doi.org/10.1063/1.1843591 [ ] .
j. berges , s. borsnyi , http://link.aps.org/doi/10.1103/physrevd.74.045022[range of validity of transport equations ] , phys .
d 74 ( 4 ) ( 2006 ) 045022 . http://arxiv.org/abs/hep-ph/0512155 [ ] , http://dx.doi.org/10.1103/physrevd.74.045022 [ ] . http://link.aps.org/doi/10.1103/physrevd.74.045022 t. arai , http://link.aps.org/doi/10.1103/physrevd.88.064029[effective potential and goldstone bosons in de sitter space ] , phys .
d 88 ( 6 ) ( 2013 ) 064029 .
http://arxiv.org/abs/1304.5631 [ ] , http://dx.doi.org/10.1103/physrevd.88.064029 [ ] .
http://link.aps.org/doi/10.1103/physrevd.88.064029 a. pilaftsis , d. teresi , http://arxiv.org/abs/1502.07986[symmetry improved 2pi effective action and the infrared divergences of the standard model ] .
http://arxiv.org/abs/1502.07986 [ ] .
s. coleman , http://link.aps.org/doi/10.1103/physrevd.15.2929[fate of the false vacuum : semiclassical theory ] , phys .
d 15 ( 1977 ) 29292936 . http://dx.doi.org/10.1103/physrevd.15.2929 [ ] .
m. beneke , p. moch , http://link.aps.org/doi/10.1103/physrevd.87.064018[on `` dynamical mass '' generation in euclidean de sitter space ] , phys .
d 87 ( 2013 ) 064018 .
http://arxiv.org/abs/1212.3058 [ ] , http://dx.doi.org/10.1103/physrevd.87.064018 [ ] .
a. malbouisson , r. portugal , n. svaiter , http://www.sciencedirect.com/science/article/pii/s0378437100005872[a non - perturbative solution of the zero - dimensional lambda phi^4 field theory ] , physica a 292 ( 1 - 4 ) ( 1999 ) 11 .
http://arxiv.org/abs/hep-th/9909175 [ ] , http://dx.doi.org/10.1016/s0378-4371(00)00587-2 [ ] . http://www.sciencedirect.com/science/article/pii/s0378437100005872 c. m. bender , s. a. orszag , https://www.springer.com/mathematics/analysis/book/978-0-387-98931-0[advanced mathematical methods for scientists and engineers i ] , springer - verlag , new york , 1999 . https://www.springer.com/mathematics/analysis/book/978-0-387-98931-0 j. ellis , e. gardi , m. karliner , m. a. samuel , pad approximants , borel transforms and renormalons : the bjorken sum rule as a case study , phys .
b 366 ( 1 - 4 ) ( 1996 ) 268275 . http://arxiv.org/abs/hep-ph/9509312 [ ] , http://dx.doi.org/10.1016/0370-2693(95)01326-1 [ ] .
m. beneke , http://linkinghub.elsevier.com/retrieve/pii/s0370157398001306[renormalons ] , physics reports 317 ( 1 - 2 ) ( 1999 ) 1142 .
http://dx.doi.org/10.1016/s0370-1573(98)00130-6 [ ] .
g. v. dunne , m. nsal , http://link.aps.org/doi/10.1103/physrevd.89.105009[uniform wkb , multi - instantons , and resurgent trans - series ] , physical review d 89 ( 10 ) .
http://dx.doi.org/10.1103/physrevd.89.105009 [ ] .
http://link.aps.org/doi/10.1103/physrevd.89.105009 g. basar , g. v. dunne , m. nsal , http://link.springer.com/10.1007/jhep10(2013)041[resurgence theory , ghost - instantons , and analytic continuation of path integrals ] , journal of high energy physics 2013 ( 10 ) .
http://arxiv.org/abs/1308.1108 [ ] , http://dx.doi.org/10.1007/jhep10(2013)041 [ ] .
u. kraemmer , a. rebhan , http://arxiv.org/abs/hep-ph/0310337[advances in perturbative thermal field theory ] , rep .
67 ( 3 ) ( 2004 ) 351431 .
http://arxiv.org/abs/hep-ph/0310337 [ ] , http://dx.doi.org/10.1088/0034-4885/67/3/r05 [ ] .
a. pilaftsis , d. teresi , http://linkinghub.elsevier.com/retrieve/pii/s0550321313003180[symmetry improved cjt effective action ] , nucl .
b 874 ( 2 ) ( 2013 ) 31 . http://arxiv.org/abs/1305.3221 [ ] , http://dx.doi.org/10.1016/j.nuclphysb.2013.06.004 [ ] . http://linkinghub.elsevier.com/retrieve/pii/s0550321313003180 h. kleinert , http://linkinghub.elsevier.com/retrieve/pii/s000349169895789x[systematic improvement of hartree - fock - bogoliubov approximation with exponentially fast convergence from variational perturbation theory ] , annals of physics 266 ( 1 ) ( 1998 ) 135161 . http://dx.doi.org/10.1006/aphy.1998.5789 [ ] . | approximations based on two - particle irreducible ( 2pi ) effective actions ( also known as @xmath0-derivable , cornwall - jackiw - tomboulis or luttinger - ward functionals depending on context ) have been widely used in condensed matter and non - equilibrium quantum / statistical field theory because this formalism gives a robust , self - consistent , non - perturbative and systematically improvable approach which avoids problems with secular time evolution .
the strengths of 2pi approximations are often described in terms of a selective resummation of feynman diagrams to infinite order
. however , the feynman diagram series is asymptotic and summation is at best a dangerous procedure .
here we show that , at least in the context of a toy model where exact results are available , the true strength of 2pi approximations derives from their self - consistency rather than any resummation .
this self - consistency allows truncated 2pi approximations to capture the branch points of physical amplitudes where adjustments of coupling constants can trigger an instability of the vacuum . this , in effect , turns dyson s argument for the failure of perturbation theory on its head . as a result
we find that 2pi approximations perform better than pad approximation and are competitive with borel - pad resummation .
finally , we introduce a hybrid 2pi - pad method . two particle irreducible effective action , resummation , nonperturbative , quantum field theory , borel summation , pad approximant arxiv : 1503.08664 81t10 , 81q20 , 81q15 , 65b10 |
a major goal of the cern large hadron collider ( lhc ) is to test the idea of weak scale supersymmetry ( susy)@xcite , wherein superpartners of the standard model ( sm ) particles have masses of the order of 1 tev .
the susy searches by the atlas and cms collaborations have reported no signal beyond sm expectations @xcite in @xmath30 fb@xmath15 of data . interpreting their results within the msugra / cmssm model @xcite , atlas and cms exclude roughly the mass range @xmath31 tev for @xmath32 , and @xmath33 gev in the case where @xmath34 .
tev , giving a gluino mass limit of about @xmath35 gev ; this limit suffers further weakening for decoupling scalars : see @xcite . ]
this reach will soon be extended since each experiment now has @xmath36 fb@xmath15 of data collected .
analysis of this extended data sample is eagerly anticipated by the hep community . within a large class of susy models
, it is expected that pair production of strongly interacting sparticles@xmath37 , @xmath38 and @xmath39 production constitutes the dominant susy production cross sections@xcite .
the gluinos and squarks are then expected to decay through a ( possibly lengthy ) cascade to lighter sparticles plus sm particles , until the decay chain terminates in the ( stable ) lightest susy particle ( lsp ) @xcite .
the lsp is expected from cosmological arguments to be a massive , neutral , weakly interacting particle ( such as the lightest neutralino @xmath40 ) and so does not deposit energy in the experimental apparatus , giving rise to the classic missing transverse energy ( @xmath13 ) signature .
thus , gluino and squark pair production followed by cascade decays is expected to give rise to final states containing multiple isolated leptons , multiple jets and @xmath13 @xcite .
while weak scale supersymmetric models are theoretically very compelling , they do suffer from a variety of problems , including 1 . the susy flavor problem , 2 .
the susy @xmath0 problem , 3 . the gravitino problem , and 4 .
the danger of too rapid proton decay in susy grand unified theories ( guts ) .
all four of these problems are greatly ameliorated if not solved by the decoupling solution , wherein first and second generation sfermion masses are pushed into the multi - tev regime or even beyond .
naturalness may be maintained in models wherein sparticles that couple directly to the higgs sector
the third generation scalars and electroweak - inos remain at or below the tev scale@xcite .
also , in many susy models , it is expected that gaugino mass parameters unify at the gut scale , in parallel with unification of gauge couplings .
renormalization group running effects result in weak scale gaugino masses occurring in the approximate ratio @xmath41 .
we would thus expect the physical gluino @xmath42 , the wino - like chargino @xmath8 and the bino - like neutralino @xmath40 to be found with roughly the same mass ratio , provided the superpotential @xmath43-parameter @xmath44 .
consequently , in models with gaugino mass unification , the experimental bounds on the gluino mass impose severe constraints on chargino and neutralino masses .
current analyses do not put independent constraints on the electroweak - ino masses if the gaugino mass unification condition is dropped @xcite .
moreover , the relative strengths of signals in various multilepton topologies ( as well as the gluino mass reach if the parent - daughter mass difference is sufficiently small ) depend sensitively on the @xmath45 and/or @xmath46 mass differences .
finally , an independent discovery of directly produced charginos and neutralinos is essential to elucidate the supersymmetry origin of any excess in the well - studied multilepton plus multijet plus @xmath13 channel at the lhc .
it is therefore interesting and relevant to find ways to discover charginos and neutralinos independently of gluinos .
another point is important to note : as we push the gluino mass to larger values , convolution of the @xmath37 subprocess cross sections with parton distribution functions ( pdfs ) requires sampling higher and higher values of parton fractional momentum @xmath47 . for such high values of @xmath47
, the parton - parton luminosity is sharply falling . at some point
we expect that , despite being strongly - produced , gluino pair production will no longer dominate over electroweak - ino pair production , since these latter reactions will sample the pdfs at much lower values of @xmath47 if electroweak - inos are significantly lighter than gluinos . to illustrate this , we plot in fig .
[ fig : xsecs ] the @xmath37 , @xmath48 and @xmath49 production cross sections in pb at lhc with @xmath50 collisions at @xmath4 tev .
our results are in nlo qcd from the program prospino@xcite .
we take @xmath51 tev for the first and second generations , in accord with a decoupling solution to the above - mentioned pathologies and , for simplicity , assume universal gaugino masses at the gut scale . from fig .
[ fig : xsecs ] , we see that gluino - pair production is dominant for @xmath52 gev . for higher values of @xmath53 ,
@xmath48 production is dominant , followed by @xmath49 production ( the reaction @xmath54 has lower cross section , and @xmath6 , @xmath55 production occurs via the unsuppressed isotriplet @xmath56 gauge coupling , whereas the @xmath57 coupling is strongly suppressed because it arises only due to the subdominant higgsino content of the wino - like chargino and the bino - like neutralino the @xmath11-bino - wino coupling is forbidden by gauge invariance .
] as can be seen _
e.g. _ in fig . 12.23 of ref .
@xcite ) . for lhc with @xmath58 tev ,
@xmath37 production remains dominant up to @xmath59 tev if squarks are very heavy . since atlas and cms already exclude @xmath33 gev when @xmath60 is large it may prove fruitful to probe electroweak gaugino pair production in the 2011 data but most of all in the 2012 lhc run .
this was recognized early on in @xcite and also more recently in in @xcite . recognizing that the stability of the higgs sector also requires sub - tev top squarks
, we also show the cross section for top squark pair production for @xmath61 by the dotted line .
in other words , for the dotted line , the graph is plotted versus @xmath62 . if other third generation squarks are also light , their pair production cross sections are also given by the dotted line with the understanding that the label on the horizontal axis is the corresponding squark mass . ] in fig .
[ fig : xsecs ] .
we see that this cross section also drops off rapidly with the top squark mass .
unless top squarks are exceptionally light ( with masses of order @xmath63 or smaller , and certainly much smaller than @xmath53 ) , electroweak - ino production remains the dominant mechanism .
let us next examine the signatures resulting from @xmath55 production . if @xmath64 , the well - known trilepton signal provides a golden signature for chargino - neutralino production@xcite provided only that the branching fraction for neutralino decay is not unduly suppressed@xcite .
the two - body chargino decay @xmath65 is expected to dominate for @xmath66 , while the two - body decay @xmath7 dominates for @xmath67 . for even higher values of @xmath68 ,
i.e. @xmath69 , the decay mode @xmath70 turns on and dominates .
this is illustrated in fig .
[ fig : bfs ] , where we show the @xmath6 branching fractions versus @xmath68 for a msugra model line with @xmath71 tev , @xmath72 , @xmath73 and @xmath74 .
we vary @xmath75 to obtain the variation in @xmath68 . in this case , @xmath76 is kinematically allowed for 175 gev @xmath77 gev , which corresponds to gluino masses in the interval 600 gev @xmath78 gev .
thus , in this mass range , we expect the single reaction @xmath79 followed by @xmath65 and @xmath7 to be the dominant susy production and decay process at lhc7 for models with full gaugino mass unification .
the endpoints of this interval can shift up or down in non - universal mass scenarios .
we begin by examining the viability of the reaction @xmath81 for susy discovery at lhc7 , focusing on the case where both @xmath82 and @xmath11 decay leptonically , resulting in clean trilepton events .
it is worth mentioning that the trilepton signal from the decay @xmath83 where a pair of opposite - sign same - flavor ( os / sf ) dileptons reconstruct the @xmath82 mass has generally been regarded as unobservable because of large sm background from @xmath19 production .
the case where the @xmath11 decays hadronically will be discussed in section [ sec : lljj ] . for our lhc7 event generation
, we use the event generator isajet 7.79@xcite for signal reactions , while for the simulation of the background events , we use alpgen@xcite and madgraph@xcite to compute the hard scattering events and pythia @xcite for the subsequent showering and hadronization . in our simulation , we include the following backgrounds for the @xmath84 signal : @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 and @xmath94 . for @xmath85 , @xmath95 , @xmath96 , @xmath97 and @xmath98
we include the full matrix elements for at least two real parton emissions and use the mlm matching algorithm to avoid double counting . for @xmath19 production
we include the full matrix elements for the @xmath99 process @xmath100 .
k - factors for both signal and background ( bg ) are included and are computed using prospino@xcite and mcfm@xcite , respectively . in our calculations , we employ a toy detector simulation with calorimeter cell size @xmath101 and @xmath102 .
the hcal ( hadronic calorimetry ) energy resolution is taken to be @xmath103 for @xmath104 and fcal ( forward calorimetry ) is @xmath105 for @xmath106 , where the two terms are combined in quadrature .
the ecal ( electromagnetic calorimetry ) energy resolution is assumed to be @xmath107 . in all these , @xmath108 is the energy in gev units .
we use the cone - type isajet @xcite jet - finding algorithm to group the hadronic final states into jets .
jets and isolated lepton are defined as follows : jets are hadronic clusters with @xmath109 , @xmath110 and @xmath111 gev . electrons and muons are considered isolated if they have @xmath112 , @xmath113 gev with visible activity within a cone of @xmath114 about the lepton direction , @xmath115 $ ] gev .
we identify hadronic clusters as @xmath116-jets if they contain a @xmath117 hadron with @xmath118 15 gev , @xmath119 3.0 and @xmath120 0.5 .
we assume a tagging efficiency of 60@xmath121 and light quark and gluon jets can be mis - tagged as a @xmath116-jet with a probability 1/150 for @xmath122 100 gev , 1/50 for @xmath123 250 gev , with a linear interpolation for 100 gev @xmath124 250 gev .
next , we invoke the following pre - selection cuts on our signal and background event samples to extract those with a @xmath125 topology : * * @xmath126 ( to aid in vetoing @xmath85 background ) , * 3 isolated leptons with @xmath127 gev and * @xmath128 gev , where two of the leptons in the event must form an os / sf pair .
if more than one os / sf pairing is possible , the pair which minimizes @xmath129 is chosen .
the remaining lepton is labeled @xmath130 . in fig .
[ fig : met ] we show the @xmath13 and transverse mass ( @xmath131 ) distributions for the signal and the sm bg after the pre - selection cuts have been applied .
the signal point has @xmath132 gev , @xmath133 gev and @xmath134 gev and we only consider @xmath55 production . due to its relatively light parent mass scale , the signal presents a soft @xmath13 spectrum , barely visible above the sm background .
this is in strong contrast with events from production of the much heavier gluinos or squarks , where the cascade decays to the lsp result in a usually much harder @xmath13 spectrum .
therefore , the usual @xmath13 plus jets / leptons searches ( optimized to look for strongly produced gluinos and squarks ) are insensitive to the @xmath135 signal . as seen in the upper frame of fig .
[ fig : met ] , after the pre - selection cuts the bg is dominated by @xmath88 production at low @xmath13 and by @xmath19 production for @xmath136 gev .
the transverse mass @xmath137 from @xmath138 , shown in the lower frame of fig .
[ fig : met ] , falls sharply beyond the expected jacobian peak at @xmath139 . in constrast ,
the corresponding signal distribution from @xmath55 production extends to considerably larger values due to the presence of the two neutralinos in the final state .
therefore , a @xmath140 cut is extremely efficient to suppress the @xmath19 background .
this is seen in the lower frame of fig .
[ fig : met ] , where the signal distribution clearly stands out for @xmath141 gev .
however , since a precise prediction for the @xmath142 tail from @xmath19 events requires a full detector simulation or data - driven estimates , we define a conservative signal region requiring : * @xmath143 gev , * @xmath144 gev .
the bg cross sections from the dominant sm processes after each of the cuts mentioned above , together with the corresponding cross sections for the representative signal point with @xmath132 gev , @xmath145 gev and @xmath134 gev , are shown in table [ tab : xsecs ] .
we stress that the signal shown in fig .
[ fig : met ] and listed in table [ tab : xsecs ] comes exclusively from @xmath55 production .
depending on the sparticle spectrum , the actual signal may be larger if heavier electroweak - inos are also accessible , or if gluino and/or squark pair production followed by their cascade decays to the @xmath19 final state is sizeable .
nonetheless , a trilepton signal would be visible with an integrated luminosity of @xmath146 fb@xmath15 at lhc7 even if light electroweak - inos are the _ only _ susy particles being produced .
.number of events generated , total cross section and cross section after cuts for the dominant backgrounds in the trilepton channel and for the signal .
all cross sections are in fb and the signal is from just @xmath55 production with @xmath147 gev , @xmath148 gev and @xmath134 gev .
the total bg values include all processes listed in the text , including the subdominant ones not shown in the table .
[ tab : xsecs ] [ cols="<,>,>,>,>,>,>,>",options="header " , ] in fig .
[ fig : mll ] , we show the @xmath149 distribution for signal and background after all the above cuts have been applied .
the dominant backgrounds displayed are @xmath85 and @xmath95 ( including @xmath150 ) . due to the @xmath13 cut
, the remaining @xmath95 contribution comes mostly from @xmath150 , with @xmath151s decaying leptonically .
therefore all @xmath95 events have @xmath152 .
for these dominant backgrounds @xmath85 , @xmath150 etc. we expect nearly equal contributions of opposite - flavor dileptons ( of ) : @xmath153 pairs , while signal is all in the sf dilepton channel .
hence the of distribution can serve as a background normalization . as seen in fig .
[ fig : mll ] , the signal is visible over the @xmath85 distribution at @xmath154 .
width smearing in the real @xmath82 emission , so all @xmath155 fall exactly at @xmath156 . ]
therefore , after applying the cuts listed above , we also require : @xmath128 gev .
as shown in table [ tab : xsecs2 ] , after the @xmath149 cut has been included , the bg is almost entirely given by @xmath85 , which contribution can be estimated using the opposite - flavor dilepton invariant mass , as mentioned above . however , although the dilepton signal rate is considerably superior to the trilepton case , the signal still is significantly below the background . in fig .
[ fig : sig-2l ] we plot the signal significance after the above cuts have been applied for various integrated luminosity values versus @xmath63 . as in fig .
[ fig : sig-3l ] we assume a msugra line with @xmath71 tev , @xmath72 , @xmath73 and @xmath74 . we see immediately that the significance in the dilepton channel is almost half of the significance in the trilepton channel , shown in fig .
[ fig : sig-3l ] .
nevertheless , corroborative evidence at the @xmath157 level is expected over almost the entire kinematically allowed range for an integrated luminosity of 20 - 30 fb@xmath15 .
in this paper , we have pointed out that for a class of susy models with decoupled matter scalars , @xmath158 and gluino masses above @xmath159500 gev , electroweak production of @xmath55 dominates the susy production rate at lhc7 .
we have examined the case where @xmath160 , for which we expect the two - body decay modes @xmath161 and @xmath65 to dominate , leading to rather simple final state topologies including @xmath162 ( trileptons ) and @xmath163 ( dilepton plus jets ) .
evaluation of the trilepton signal against sm backgrounds shows that the susy signal should be observable with a @xmath164 significance at lhc7 up to @xmath165 gev ( depending on @xmath16 ) , for an integrated luminosity of 30 fb@xmath15 . in models with gaugino mass unification
, this corresponds to a range in gluino masses of @xmath166 gev .
moreover , we find that for most of this region a @xmath167 excess is expected in the 5 fb@xmath15 data sample that has already been accumulated . assuming 30 fb@xmath15 of integrated luminosity at lhc7
, the trilepton channel will be competitive in reach with the canonical multijet plus @xmath13 search in models with unified gaugino mass parameters .
if a signal is seen in the trilepton channel , a @xmath168 confirmatory signal is also expected in the dilepton plus jets channel for most of the parameter space , thus making a stronger case for the @xmath135 signal .
most importantly , the simultaneous presence of these signals will point to the susy origin of any new physics that might be discovered in the 2012 run .
this work was supported in part by the u.s .
department of energy under grant nos .
de - fg02 - 04er41305 , de - fg02 - 04er41291 and de - fg02 - 95er40896 , by the in2p3 of france under contract pics fr - usa no .
5872 and by the fundaco de apoio pesquisa do estado de so paulo ( fapesp ) .
99 for a review of susy , see h. baer and x. tata , _ weak scale supersymmetry : from superfields to scattering events _ , ( cambridge university press , 2006 ) .
atlas collaboration , g. aad _
et al . _ , arxiv:1109.6572 , arxiv:1109.6606 , jhep 11 ( 2011 ) 99 , and arxiv:1110.6189 ;
see also + https://twiki.cern.ch/twiki/bin/view/atlaspublic/supersymmetrypublicresults cms collaboration , s. chatrchyan _
_ , _ phys .
lett . _ * 107 * ( 2011 ) 221804 , cms - pas - sus-11 - 008 ; see also https://twiki.cern.ch/twiki/bin/view/cmspublic/physicsresultssus a. chamseddine , r. arnowitt and p. nath , phys .
rev . lett . * 49 * , 970 ( 1982 ) ; r. barbieri , s. ferrara and c. savoy , phys .
b*119 * , 343 ( 1982 ) ; n. ohta , prog .
. phys . * 70 * , 542 ( 1983 ) ; l. j. hall , j. lykken and s. weinberg , phys .
* d27 * , 2359 ( 1983 ) ; for a review , see _ e.g. _ r. arnowitt and p. nath , arxiv:0912:2273 ( 2009 ) .
atlas collaboration , atlas - conf-2011 - 098 .
h. baer , c. h. chen , f. paige and x. tata , _ phys . rev . _ * d 52 * ( 1995 ) 2746 .
h. baer , c. h. chen , f. paige and x. tata , _ phys .
rev . _ * d 53 * ( 1996 ) 6241 .
h. baer , v. barger , d. karatas and x. tata , _ phys .
_ * d 36 * ( 1987 ) 96 ; h. baer , r. m. barnett , m. drees , j. f. gunion , h. e. haber , d. l. karatas and x. r. tata , int .
j. mod .
phys . a * 2 * , 1131 ( 1987 ) ; h. baer , a. bartl , d. karatas , w. majerotto and x. tata , _ int .
j. mod .
phys . _ * a 4 * ( 1989 ) 4111 ; h. baer , x. tata and j. woodside , _ phys .
* d 42 * ( 1990 ) 1568 ; for earlier work on sparticle decays to just gauginos , see h. baer , j. ellis , g. gelmini , d. v. nanopoulos and x. tata , _ phys .
* b 161 * ( 1985 ) 175 ; g. gamberini , _
z. physik _ * c 30 * ( 1986 ) 605 ; h. baer and e. berger , _ phys . rev .
_ * d 34 * ( 1986 ) 1361 .
h. baer , x. tata and j. woodside , _ phys .
* d 41 * ( 1990 ) 906 and _ phys .
* d 45 * ( 1992 ) 142 .
m. dine , a. kagan and s. samuel , _ phys .
* b 243 * ( 1990 ) 250 ; n. arkani - hamed and h. murayama , _ phys .
* d 56 * ( 1997 ) 6733 .
m. drees , _ phys .
_ * d 33 * ( 1986 ) 1468 ; s. dimopoulos and g. giudice , _ phys .
* b 357 * ( 1995 ) 573 ; a. pomarol and d. tomassini , _ nucl .
* b 466 * ( 1996 ) 588 ; a. cohen , d. b. kaplan and a. nelson , _ phys .
* b 388 * ( 1996 ) 588 ; s. sekmen , _ et al .
_ , arxiv:1109.5119 [ hep - ph ] .
w. beenakker , r. hopker , m. spira , hep - ph/9611232 ( 1996 ) .
s. mrenna [ for the cms collaboration ] , arxiv:1110.4078 [ hep - ph ] .
h. baer , v. barger , a. lessa , w. sreethawong and x. tata , arxiv:1201.2949 .
h. baer , c. h. chen , f. paige and x. tata , _ phys .
* d 50 * ( 1994 ) 4508 .
h. baer and x. tata , _ phys .
rev . _ * d 47 * ( 1993 ) 2739 .
f. paige , s. protopopescu , h. baer and x. tata , hep - ph/0312045 . m. mangano , m. moretti , f. piccinini , r. pittau and a. polosa , _ j. high energy phys .
_ * 0307 * ( 2003 ) 001
. j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , _ j. high energy phys . _
* 1106 * ( 2011 ) 128 .
t. sjostrand , s. mrenna and p. skands , _ j. high energy phys . _
* 0605 * ( 2006 ) 026 .
mcfm , by j. campbell and r. k. ellis .
see r. k. ellis , nucl .
* 160 * ( 2006 ) 170 .
h. baer , v. barger and a. mustafayev , arxiv:1112.3017 . h. baer , v. barger , a. lessa and x. tata , arxiv:1112.3044 [ hep - ph ] . | lhc searches for supersymmetry currently focus on strongly produced sparticles , which are copiously produced if gluinos and squarks have masses of a few hundred gev . however , in supersymmetric models with heavy scalars , as favored by the decoupling solution to the susy flavor and @xmath0 problems , and @xmath1 gev as indicated by recent lhc results , chargino
neutralino @xmath2 ) production is the dominant cross section for @xmath3 at lhc with @xmath4 tev ( lhc7 ) .
furthermore , if @xmath5 , then @xmath6 dominantly decays via @xmath7 , while @xmath8 decays via @xmath9 .
we investigate the lhc7 reach in the @xmath10 channel ( for both leptonic and hadronic decays of the @xmath11 boson ) in models with and without the assumption of gaugino mass universality . in the case of the msugra / cmssm model with heavy squark masses ,
the lhc7 discovery reach in the @xmath12 channel becomes competetive with the reach in the canonical @xmath13 + jets channel for integrated luminosities @xmath14 fb@xmath15 .
we also present the lhc7 reach for a simplified model with arbitrary @xmath16 and @xmath17 . here
, we find a reach of up to @xmath18 gev for 10 ( 30 ) fb@xmath15 .
pacs numbers : 14.80.ly , 12.60.jv , 11.30.pb @xmath19 plus missing-@xmath20 signal + from gaugino pair production at lhc7 + howard baer@xmath21 , vernon barger@xmath22 , sabine kraml@xmath23 , andre lessa@xmath24 , + warintorn sreethawong@xmath25 and xerxes tata@xmath26 + _ @xmath25dept . of physics and astronomy , university of oklahoma , norman , ok 73019 , usa + @xmath27dept . of physics ,
university of wisconsin , madison , wi 53706 , usa + @xmath28laboratoire de physique subatomique et de cosmologie , ujf grenoble 1 , cnrs / in2p3 , inpg , 53 avenue des martyrs , f-38026 grenoble , france + @xmath29instituto de fsica , universidade de so paulo , so paulo - sp , brazil + @xmath26dept . of physics and astronomy , university of hawaii , honolulu , hi 96822 , usa + _ |
during the past decade , there has been a steady improvement in observations of deuterium to hydrogen ( d / h ) in the universe .
the measurement accuracy is now good enough to provide insight into galaxy evolution .
first of all , the primordial d / h ratio ( d / h)@xmath0 is confined to a narrow range of @xmath1(2.6 - 2.8)@xmath2 .
this range covers the analysis result of data from the _ wilkinson microwave anisotropy probe _ ( wmap ) that gives 2.60@xmath3@xmath2 @xcite or 2.75@xmath4@xmath2 @xcite , depending on the different light - element nuclear reaction rate ( also [email protected] from the wmap 5 data ; * ? ? ?
* ) , and the measurements of distant metal - poor damped ly@xmath6 systems that lead to 2.82@xmath7@xmath2 ( * ? ? ?
* see also omeara et al .
recent _ far - ultraviolet spectroscopic explorer _ ( fuse )
observations reveal that the present - day d / h abundance ( d / h)@xmath8 for interstellar matter ( ism ) in the galactic disk is surprisingly as high as @xmath9 2.31@xmath5 0.24@xmath2 after correcting for dust depletion @xcite .
some authors claim there are residual effects due to dust depletion which can account for the large scatter in the observed d / h @xcite . but the high d / h abundance of 2.2@xmath10@xmath2 in the warm neutral medium ( wnm ) of the lower galactic halo , which originates in the disk and is elevated into the halo @xcite , strongly supports the claim by @xcite . the difference between ( d / h)@xmath0 and ( d / h)@xmath8 is defined as a deuterium astration factor @xmath11=(d / h)@xmath0/(d / h)@xmath8 .
the value of @xmath11 should exceed unity because it is believed that all d is produced during big bang nucleosynthesis , and any subsequent destruction process of d occurs in the interior of stars .
thus , the value of @xmath11 is determined by the fraction of matter which has never been cycled through stars in the present ism .
accordingly , @xmath11 is closely related to the degree of heavy - element enrichment through star formation .
our expectation is that the d abundance in the ism is anti - correlated with the metallicity such that the well - enriched ( @xmath1solar metallicity ) local disk demands a high @xmath11 .
indeed , conventional galactic chemical evolution ( gce ) models have been applied to the evolution of d so far by several authors @xcite , and predict @xmath12 for the local disk @xcite . in other words , the local abundances of heavy elements such as fe or @xmath6-elements in the ism as well as long - lived stars imply @xmath13 at the least , that is equivalent to ( d / h)@xmath142.0@xmath15 . on the other hand , the high ( d / h)@xmath8 suggested by the recent observations implies a smaller @xmath11 . for the values of ( d / h)@xmath0=2.75@xmath2 and ( d / h)@xmath8=2.31@xmath2
, we obtain @xmath11=1.19 , which is incompatible with the theoretical prediction by gce models .
the d / h abundances which bridge between the primordial ( d / h)@xmath0 and the present - day ( d / h)@xmath8 provide crucial information on the evolution of d. the d / h abundance in the protosolar cloud provides crucial information at a look - back time of @xmath1 4.6 gyr .
the d / h abundances deduced from the jovian atmosphere @xcite and from the solar wind @xcite are estimated to be @xmath16 0.4 ( 0.5)@xmath2 .
this result is broadly consistent with the present - day local value .
if these values are representative of the local disk , we are led to believe that d astration involved with star formation has turned off for the last several gyr .
in fact , as the process to change the amount of d in the ism , the galactic disk involves an enrichment process through an infall of low - metallicity gas from the halo , which brings ( nearly ) primordial d , besides its destruction through star formation .
this infall is required by the gce models to reproduce the abundance distribution function ( adf ) of local disk stars ( e.g. , * ? ? ?
in addition , we cite high - velocity clouds ( hvcs ) as the evidence for the present - day infall ( e.g. , * ? ? ? * ; * ? ? ? * ) .
the infall operates in the gce models so as to enrich the ism efficiently in the early phase to suppress the number of metal - poor stars , and maintain the star formation activity by continuously suppling the gas .
this established view predicts a monotonic increase ( decrease ) in heavy elements ( d ) in the ism up to the solar abundances ( ( d / h)@xmath172.0@xmath15 ) . in this way
, the evolution of d is tightly bound to the production of metals in star formation .
but what if a significant fraction of the heavy elements in the disk does not originate from _ in situ _ star formation ( e.g. * ? ? ?
this represents a radical departure from conventional gce models and has recently been discussed in detail . here
large - scale winds from the galactic bulge , which entrain a large amount of heavy elements , enrich the disk .
the hypothesis well explains the origin of metal - rich stars in the solar neighborhood @xcite and the mechanism of the time - evolution of abundance gradient @xcite . in this scheme ,
a part of heavy elements trapped in long - lived stars as well as in the gas of the disk is not related to the local star formation but the activity of the bulge . in this paper
, we shed light on the evolution of deuterium in the framework that the galaxy disk has evolved through `` wind + infall '' , together with the renewed constraint from the star formation history in the local disk .
then we show that the combined phenomena of winds that entrain heavy elements and infall that contains primordial d will make the galaxy disk feature withstand the test of observations .
in this section , we summarize the proposed wind scenario , which is a key ingredient in the making of a heavy - element feature in the disk , based on the discussions in @xcite and @xcite , together with the updated observational results .
_ 1 . metal - rich stars in the solar neighborhood _
there exist metal - rich stars beyond the solar metallicity , the fraction of which is roughly 20% .
in fact , its presence as well as its elemental feature poses a puzzling problem .
the adf of solar neighborhood disk stars has been studied by many authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and the location of its peak at [ fe / h]@xmath18@xmath19 -0.15 has been firmly established . on the other hand
, the metal - rich end of the adf extends at least to [ fe / h]@xmath1 + 0.2 .
spectroscopic observations of elemental abundances for metal - rich disk stars @xcite have confirmed that chemical enrichment in the solar neighborhood continued unabated until [ fe / h]@xmath1 + 0.4 .
@xcite has claimed that the simultaneous reproduction of both the presence of stars with [ fe / h]+0.2 and [ fe / h]@xmath200 is hard to realize through the conventional scheme of local enrichment via low - metallicity gaseous infall from the halo , and that the presence of supersolar stars is crucial evidence for enrichment by winds from the bulge .
it should be of note that the seeming making of supersolar stars by gce models thus far ( e.g. , * ? ? ?
? * ; * ? ? ?
* ) is an end result of convolution of the adf with a dispersion of 0.1 - 0.2 dex in [ fe / h ] .
the convincing signature of wind enrichment is seen in their elemental features .
nearby metal - rich stars are represented by an upturn in [ @xmath6/fe ] ratios except for [ o / fe ] against increasing [ fe / h]@xcite .
such upturns can also be seen in other elements , such as odd elements ( na , al ) , and some iron - group elements ( ni , zn ) .
it is incorrect to attribute these upturns to metallicity - dependent sn ii yields , especially because this feature is observed with both even- and odd - numberd elements .
these features can be interpreted as a reflection of the elemental feature of winds from the bulge @xcite . since the [ @xmath6/fe ] ratios start to upturn from [ fe / h]@xmath21 , the wind enrichment is expected to occur around 5 gyr ago , corresponding to the time of sun s birth . from the above discussion on elemental feature , the galactic fountain ( e.g. , * ? ? ?
* ; * ? ? ?
* ) that may serve as a means for transporting metals to other locations in the disk is unlikely to contribute to acceleration of the chemical enrichment in the solar vicinity .
on the other hand , radial mixing of disk stars due to resonant scattering with transient spiral waves predicts that contaminants coming from the inner disk populate the solar neighborhood ( sellwood & binney 2002 ; rokar et al .
2008a , b ; schenrich & binney 2009 ; snchez - blzquez et al . 2009 ) , and allows the presence of metal - rich stars beyond the upper limit of a local chemical enrichment @xcite .
however , stars orbiting near the galactic center should exhibit a decreasing [ @xmath6/fe ] trend with increasing [ fe / h ] , as observed in the bulge @xcite .
therefore , elemental features of nearby metal - rich stars seem at odds with the prediction of radial mixing as a mechanism of producing metal - rich stars
. it should be of note that some authors claimed no upturning feature but a constant solar ratio for nearby metal - rich stars @xcite .
however , none of the existing models including radial mixing supports this result .
time evolution of the abundance gradient _
@xcite has found that the radial [ fe / h ] gradient roughly from @xmath22=4 to @xmath22=14 kpc ( @xmath22 : galactocentric distance ) has flattened out in the last several gyr with a change in a slope of @xmath23 - 0.1 dex kpc@xmath24 to @xmath1 -0.05 dex kpc@xmath24 , by comparing the steep relic gradient observed in the open clusters with that observed in the cepheids ( see also * ? ? ?
* ; * ? ? ?
chemical evolution models predict contradictory time evolution of the metallicity gradient , i.e. , a steepening @xcite or a flattening @xcite .
putting aside these contradictions , existing gce models predict a monotonic increase in abundances for each region , and the predicted gradient change is small compared with the observations .
@xcite has claimed that large - scale winds from the galactic bulge are the crucial factor for its mechanism .
their proposed scenario is ( i ) winds once set up a steep abundance gradient ( @xmath1 -0.1 dex kpc@xmath24 ) owing to the enrichment by heavy elements entrained in the winds , and ( ii ) later evolution leads to a flattening of abundance gradient through chemical evolution under an accretion of a low - metal gas from the halo .
on the other hand , new analysis of planetary nebulae supports a slight steepening of the [ o / h ] gradient with time @xcite .
this contradictory result is interesting because it is likely that winds from the bulge are not so metal - rich in o. the metal - rich bulge stars exhibit the low [ o / fe ] below the solar ratio @xcite .
abundance of the microlensed stars recently analyzed confirmed [ o / fe]@xmath1 -0.2 for metal - rich bulge stars
the reduced o yield of sn ii expected from these observations is probably due to efficient mass loss for massive stars , which prevents he and c from synthesizing into o within the metal - rich environment in the bulge @xcite .
thus , o is not a good indicator for the wind enrichment that is predicted to cause the time evolution of the abundance gradient in our scheme .
in other words , the time evolution of the [ o / h ] gradient could be possibly discussed in the conventional gce scheme as predicted by @xcite .
alternatively , it is possible that some effect by radial mixing which causes a steepening of abundance gradient @xcite is seen in the evolution of o gradient owing to little influence of the wind enrichment .
in addition , @xcite compiled a new data set of ages , [ fe / h ] , and distances for open clusters . from their samples , we make two groups : one is composed of young clusters whose ages are less than 1 gyr , and the other is of old ones whose ages range within @xmath25 gyr . here
we restrict both of samples to the ones with @xmath2614 kpc . as a result
, the least - square fitting gives a slope of -0.05 dex kpc@xmath24 for young clusters , and -0.14 dex kpc@xmath24 for old clusters , respectively .
we see a significant flattening of the [ fe / h ] gradient for new open cluster data .
_ 3 . other signatures _ in the galaxy
, there is now strong evidence for large - scale winds across the electromagnetic spectrum @xcite .
furthermore , there is tantalizing direct evidence that dust and gas from the inner galaxy have been transported to the solar neighbourhood @xcite . in standard chemical evolution models ,
the isotopes @xmath27si and @xmath28si become more abundant than @xmath29si as the disk ages . but
pre - solar grains from meteorites show evidence that the local isotopes formed in material that had experienced more nuclear synthesis than material that came after the sun s birth
. a driver of large - scale winds would be a starburst in the galactic bulge .
therefore , our scenario proposes that the galactic bulge has a complex history involving at least one starburst episode around 5 gyr ago in addition to the initial burst that gave rise to the old population @xmath30 10 gyr ago . to date
, the star formation history of the galactic bulge still remains an open question because we do not have the precise color - magnitude diagram ( cmd ) of the bulge owing to the difficulty of discrimination of bulge stars from disk stars ( e.g. , * ? ? ?
* ; * ? ? ?
there is good evidence for recent star formation in the nuclear region @xcite .
in addition , from the perspective of bar formation , one might expect some star formation activity in the region of what is now the bulge , about 3 gyr after the inner disk started to form - this would be at the time when the bar buckling occurs @xcite - it would really stir up any remaining gas in the inner disk .
it is difficult to conjecture how massive a starburst is likely to be from a viewpoint of a wind enrichment on the disk since there are too many uncertain factors to predict it . here
we take a simplified case and show that a starburst which produce @xmath31 of the bulge stars might be one candidate , by checking supply and demand of fe . for simplicity ,
the disk is here considered to be one zone represented by the solar circle .
let us take the masses of a bulge and a disk and the gas fraction of a disk as @xmath32 , @xmath33 , and 0.1 .
the imf with a slope of @xmath34= -1.05 @xcite together with the star formation producing the fraction of @xmath35 of bulge stars yields @xmath36 sne ii that eject @xmath37 of fe , some fraction of which will be , though , retained within the bulge . on the other hand , the fe abundance of gas in the disk should increase from [ fe / h]=0 to [ fe / h]=+0.3 by the wind enrichment plus the local enrichment by star fromation .
it means that the winds should entrain roughly the same amount of fe as that already present in the ism to double the fe abundance if we neglect the local fe production . in the end , the total fe mass necessary for the additional enrichment on the disk amounts to @xmath38 , equivalent to the former estimate .
if we assume that the wind enrichment is restricted to the inner disk including the solar vicinity together with a partial contribution to the winds from type ia supernovae ( sne ia ) and hypernovae that eject high fe masses , the necessary fraction of stars produced in a burst will significantly decrease like @xmath39 1% . while numerous studies @xcite agree that the bulk of the bulge formed @xmath3010 gyr , a minority population of stars as young as 5 gyr can not be ruled out , and a minor burst of star formation as recent as 5 gyr may have taken place in the bulge .
the fraction of stars produced by such a minor burst must be small like @xmath40% as implied by an above simplified estimate so as not to contradict the present cmds , but in practice , how much of stars the starburst was likely to produce remains quite unclear .
as well , the age of a burst claimed here is not definitive due to a lack of determinant factor of its age - dating .
the results from microlensed dwarf and subgiant stars in the galactic bulge showing the average age of @xmath1 7 gyr for metal - rich ( [ fe / h]@xmath30 0 ) bulge stars @xcite may be in favor of our hypothesis .
recently , some authors claim that momentum - driven winds are an efficient mechanism for a powerful wind from the central starburst @xcite .
these strong winds could realize a widespread enrichment over the disk , possibly including the outer disk region , through hydrodynamical interaction between the winds and the gaseous halo @xcite . in n - body hydrodynamic simulations
, only high - resolution models can create inhomogeneous ism sufficient for the right treatment of winds from star - forming region ( e.g. , * ? ? ? * ) .
ongoing advanced numerical calculations will certainly reveal a full view of the wind enrichment .
the main body of the galactic bulge was considered to be formed in a starburst @xmath30 10 gyr ago .
it is natural that we expect the action of ravaging winds associated with this vast star formation .
since the age of thin disk is as low as @xmath41 10 gyr ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the signature of initial wind enrichment is likely to be unseen in the thin disk stars .
instead , some fossil record should be imprinted in the formation of thick disk prior to thin disk .
indeed , recent study claims that ages of the thick disk stars are @xmath308 - 9 gyr with a span of @xmath13 gyr @xcite , which suggests a similar formation epoch between thick disk and bulge .
here we discuss compelling evidence for the contamination of a gas used for the thick disk formation by heavy elements entrained in the winds .
the adf of thick disk shows a sharp increase from [ fe / h]@xmath42 to the peak located at [ fe / h]@xmath43 .
this feature is first reported by @xcite with a small number of but complete samples , and the location of its peak is confirmed by a huge sdss data base of thick disk stars @xcite .
these observational results suggest that a material for the proto - thick disk was pre - enriched up to [ fe / h]@xmath1 -1.3 at @xmath30 10 gyr ago .
our claim is that the mechanism of pre - enrichment is attributable to the wind occurrence triggered by an initial starburst in the galactic bulge .
detailed results of not only the adf but also the evolution of [ @xmath6/fe ] calculated with the wind models will be presented in tsujimoto ( 2010 , in preparation ) .
in addition , similarity of stellar elemental features between bulge and thick disk recently revealed @xcite seems to have some connections to the wind phenomena during their overlapping formation episode . from this viewpoint , metal - weak thick disk stars can be discussed .
@xcite found that some fraction of low - metallicity stars in the range of @xmath44[fe / h]@xmath45 belong to the thick disk population . according to @xcite , addition of these stars to the adf deduced by @xcite
does not form a metal - poor tail .
rather , it produces another distinct distribution for [ fe / h]@xmath46 . in our view
, metal - weak thick disk stars are regarded as the stars that are formed prior to the significant influence of metal - rich winds from the bulge on the proto - thick disk .
we incorporate the enrichment process by large - scale winds into the standard gce model .
the basic picture is that the disk was formed through a continuous low - metal infall of material from outside the disk region , and for a specific period heavy elements carried by winds from the bulge dropped and enriched the disk .
the period of the occurrence of winds is set to be 4 - 6 gyr ago so as to reproduce the chemical features of metal - rich solar neighborhood stars as well as the observed flattening of abundance gradient in the last several gyr @xcite .
the abundances of the winds are set to be identical to the ejecta of type ii sne ( sne ii ) such as [ fe / h]=+0.4 , [ mg / fe]=+0.4 , and [ o / fe]=-0.2 .
a low o abundance in the wind is implied by the bulge abundance @xcite .
details of the models including the equations to be solved are described in @xcite .
the key model ingredients , infall rate and star formation rate ( sfr ) , are presented in the following separate subsections .
on other model inputs , their individual descriptions are simply mentioned below .
the initial mass function ( imf ) should be chosen carefully in this study since it determines the returned fraction @xmath47 , i.e. , the fraction of stellar mass that is returned to the ism , which crucially influences the evolution of d ( e.g. , * ? ? ?
we adopt a power - law mass spectrum with a slope of -1.35 , together with the mass range [ 0.05 , 50 ] .
this combination is deduced from the theoretical analysis considering two conditions given by the chemical abundances of halo stars and the mass - to - luminosity ratio in the solar neighborhood @xcite .
then , this adopted imf is combined with the nucleosynthesis yields of sne ii and type ia sne ( sne ia ) taken from @xcite . for the occurrence frequency of sne ia
, we assume that the fraction of the stars that eventually produce sne ia for @xmath48is 0.05 with the lifetime of sn ia progenitors of 0.5@xmath493 gyr @xcite .
note that the returned fraction appears in the equations to solve the evolutions of gas and abundances ( o , mg , fe ) with no instantaneous recycling approximation .
the evolution of d is described by a simple equation that is governed by the time changes of a gas fraction @xmath50 , a sfr @xmath51 , and an infall rate @xmath52 ; @xmath53 where d in the winds is neglected since the winds mostly consists of the ejecta of sne ii with no d. the d abundance in the infall is assumed to be primordial since the metallicity ( [ fe / h ] ) of the infalling gas is expected to be much lower than the present - day infall such as complex c whose metallicity is subsolar and d / h is estimated to be [email protected]@xmath2 @xcite . in the following , we will describe the key ingredients that determine the d / h evolution , together with the observational constraints to each of them . .
_ middle panel _ : the calculated star formation histories .
two model cases are considered .
one ( solid line ) involves a setup of a decreasing sfr for 0@xmath544 gyr so as to be compatible with the observed tendency ( dashed line ) derived by @xcite .
the other ( dotted line ) is the one with an assumption of a constant sfr coeficient .
both models are normalized to become the present - day stellar surface density of 35.5 pc@xmath55 .
the observed sfr at the present time is deduced from the mass growth of the stellar disk in the last 1 gyr ( see * ? ? ?
_ lower panel _ : the evolution of a gas fraction calculated with the decreasing sfr model , together with the present - day value .
] for the infall rate , we adopt the formula that is proportional to @xmath56 with a timescale of infall @xmath57 .
the value of @xmath57 is assumed to be 5 gyr so as to reproduce the observed feature of metal - poor tail of the adf @xcite . to compare with the observations , we give an infall rate in units of yr@xmath24kpc@xmath55 , that is shown on the upper panel of figure 1 . here we normalize this function so that its integration over 14 gyr sums up to today s total surface density in the local disk . today
s density is estimated to be 42 pc@xmath55 from the surface density of the stellar disk ( 35.5 pc@xmath55 ; * ? ? ?
* ; * ? ? ?
* ) and of the interstellar gas ( 6.5 pc@xmath55 ; * ? ? ?
this together with our infall function predicts the present rate of 0.00185 yr@xmath24kpc@xmath55 . on the other hand
, complex c gives an associated mass flow of 0.00067 - 0.003 yr@xmath24kpc@xmath55 @xcite . in the figure
, we adopt these values as an observationally implied range .
however , the average rate given by the entire hvc system may be higher by a factor of @xmath58 as predicted so far ( e.g. , * ? ? ?
* ; * ? ? ?
still , our prediction for the present - day rate falls within the observationally possible range .
furthermore , it should be of note that a significant number of ionized high - velocity gas detected in the vicinity of the galaxy must be a factor to increase the present - day rate ( e.g. , * ? ? ?
* ; * ? ? ?
the sfr is assumed to be proportional to the gas fraction @xmath50 with a coefficient @xmath59 ( gyr@xmath24 ) . in the standard gce models ,
the value of @xmath59 is set to be constant over the age of a galaxy . on the other hand ,
recent several works reveal the star formation history in the local disk especially for the last several gyr @xcite .
the common feature claimed by them is that the sfr has been clearly declining with time for the last several gyr . for the purpose of its reproduction ,
we introduce the time - dependent @xmath59 for the recent sfr so as to obey the observed sfr-@xmath60 relation . using the relation found by @xcite giving a decreasing rate of @xmath1 -0.8 gyr@xmath24
, we assume @xmath59=0.6 for @xmath61 gyr and @xmath59=0.6@xmath62 for 10@xmath63 14 gyr .
the middle panel of figure 1 shows the history of sfr where its normalization is done so that the present - day stellar density beomes 35.5 pc@xmath55 . for reference ,
the case with a constant @xmath59 ( = 0.6 ) over the full age is indicated by dotted line , demonstrating that a constant @xmath59 model can not reproduce the observed sfr over the last gyr .
the present sfr is deduced from the mass growth of stellar density over the last 1 gyr . from the result of @xcite with an associated uncertainty
, it is estimated to be 0.48 - 1.1 pc@xmath55 gyr@xmath24 , which covers the results obtained from other studies including @xcite . on the other hand ,
the predicted sfr is 0.82 pc@xmath55 gyr@xmath24 .
we see a good agreement with the observations .
finally , we see the evolution of gas fraction @xmath50 as a result of adopted infall rate and sfr since @xmath50 is a determinant factor to control the d / h abundance ( the lower panel of fig . 1 )
. the predicted @xmath50 is compared with the observed value at the present time .
an error in the observed @xmath50 mainly originates from a large uncertainty in the surface density of the ism .
here we adopt its range from 6.5 pc@xmath55@xcite to 13 pc@xmath55@xcite , together with the stellar density of 35.5 pc@xmath55 . a gradual increase in @xmath50 over the last 4 gyr while the sfr decreases seems at odds with the schmidt - kennicut law .
however , a span in the sfr against a small range of @xmath50 in our result is compatible with the intrinsic scatter in the observed relation between the sfr and the surface density of the ism @xcite .
what primarily induces this implied low star formation efficiency ( sfe ) irrespective of the steady supply of h i gas remains unclear , including the possible association with the low sfe deviated from the standard law in a low gas density environment such as 5 - 10 pc@xmath55 observed for faint dwarf galaxies @xcite and low surface brightness ( lsb ) galaxies @xcite . interestingly , lsb galaxies exhibit about one - order change in sfr while those have similar gas densities .
in addition , the schmidt - kennicutt law is valid only in a large scale such as @xmath30 1 kpc @xcite , whereas the local sfr observationally covers tens of pc .
we proceed to the next issue on chemical features of the local disk .
recall that we need to discuss the d / h evolution in the framework that gives a fully consistent explanation to the observed quantities on heavy elements such as fe and @xmath6-elements .
the left panel of figure 2 shows the successful reproduction of the adf by our model .
a stall of chemical enrichment owing to the decreasing sfr over the last gyrs is compensated by the effective wind enrichment . in other words ,
the activity of local star formation is not sufficient to provide all of fe contained in the existing long - lived stars .
this is shown by the model case without the winds ( dashed - line ) , which is skewed to the lower - metallicty , compared with the observations .
this wind enrichment imprints its unique record on the elemental ratios of metal - rich stars .
the winds are expected to exhibit the high [ mg / fe ] and low [ o / fe ] ratios as mentioned in the model description . accordingly , as a result of their reflections , our model predicts an upturning feature of [ mg / fe ] for metal - rich stars through a rapid increase - gradual decrease loop in the [ mg / fe]-[fe / h ] correlation , which is in good agreement with the observed data ( the right upper panel ) .
the different [ o / fe ] evolution can be considered in the same theoretical scheme with an assumption of a low [ o / fe ] in the winds ( the right lower panel ) .
now we are all set to start on the discussion of d / h evolution .
the predicted time - evolution of d / h is shown in figure 3 . here the primordial ( d / h)@xmath0 is set to be 2.75@xmath2 @xcite . from its initial value ,
the d / h abundance starts to decrease with time , and subsequently its change turns to a gradual increase 4 gyr ago .
this turnover results from the supply of primordial d by an infall that overwhelms the destruction of d through star formation owing to its low rate .
then , it finally enables to achieve a present - day high d / h abundance of 2.3@xmath2 , fully consistent with the observed value @xcite , after passing d / h @xmath64 equivalent to the protosolar value @xcite around 4.6 gyr ago . for reference , the case with a constant @xmath59 ( = 0.6 ) over the full age is indicated by dashed line .
our safe conclusion is that the d / h abundance has little evolution during the last several gyr , as if the d astration through star formation has turned off .
it s because the metallicity of an infall would change in accordance with the cosmic evolution of damped ly@xmath6 systems @xcite , and eventually become subsolar with a smaller d than the primordial value like complex c @xcite .
this view will lead to a flattening of d / h evolution rather than its gradual increase .
in this section , we briefly review the uncertainties in two key observational constraints that largely influence the results presented here . as discussed below , neither of two has reached a firm consensus yet , thus further studies to better understand these issues , observationally and theoretically , should be still awaited . in our models , we adopt a sharply decreasing sfr which changes a factor of @xmath65 in the last 4 gyr deduced by @xcite . their result is broadly consistent with other several works using different methods @xcite . however , at least two studies present a different picture for the recent sfr . @xcite
derive the star formation history , using a chromospheric age distribution of dwarf stars , and their major finding is that the disk has experienced a few bursts in the past , in which one of enhanced episodes of star formation occurred during the last 1 gyr .
on the other hand , from a new reduction of hipparcos data combined with updated isochrones , @xcite obtain the deceasing sfr from the past to the present but at a very small rate . with no doubt , major progress on this issue will come from a huge database brought by gaia in the next decade .
the controversial situation on the representative value of ( d / h)@xmath8 is nicely reviewed in ( * ? ? ? * and references therein ) . although most claims seem inclined to be above ( d / h)@xmath8 @xmath66 , that is beyond the predictions by the conventional gce models , very low values such as ( d / h)@xmath671.5@xmath2 of the local bubble value or @xmath10.98@xmath2 of the highest column density regions @xcite can not be still discarded .
the difficulty lies in the interpretation of a widespread variation in the observed gas - phase d / h ratios among local ism regions .
it seems reasonable to claim that the ism exhibiting a high d / h such as @xmath12.2 - 2.3@xmath2 is a result of incompletely mixed infall , i.e. , a strong influence by a low - metal infall ( e.g. , * ? ? ?
* ) since complex c has a similar d / h of 2.2@xmath2 .
however , it is stemmed by a counterargument that their o abundance is not so low as that expected in an infalling gas @xcite .
recently , analysis of linsky s data by a bayesian approach gives the best estimate of ( d / h)@xmath672.0@xmath2 @xcite .
a reduction of @xmath68 draws the observed ( d / h)@xmath8 close to the compatibility with some of gce models .
however , considering an uncertainty in the adopted probability distribution for depletion as well as an indication of high d in wnm @xcite , further studies with more data is still necessary for the settlement of a true ( d / h)@xmath8 . besides , recall that the observed d / h abundance only gives a lower limit of ( d / h)@xmath8 . to date , a high ( d / h)@xmath8 as proposed by @xcite is viewed with a skeptical eye owing to an incompatibility with the theoretical gce scheme .
here we propose a new path to end up with ( d / h)@xmath692.3@xmath2 .
accordingly , we claim a need to have a fresh look at those observed high d / h , bearing in mind that they are within the predictions of a renewed gce picture . ,
[ mg / fe ] , and d / h . _ left panels _ : solid and dashed lines represent the results by the model with the decreasing sfr and with a constant sfr coefficient , respectively , under the assumption of no winds . _ right panels _ : same as in left panels but for an introduction of the winds . ]
finally , let s review each role of the sharply decreasing sfr and the metal - rich winds that are the main drivers of the successful results through the comparison of two models
. left panels of figure 4 show the time evolution of [ fe / h ] , [ mg / fe ] , and d / h calculated with the model in which the sharply decreasing sfr for the last 4 gyr is incorporated into the gce scheme .
these witness that an inactive sf period makes the d / h evolution compatible with the observation , while leaving the problem of an incomplete chemical enrichment .
then , the satisfactory enrichment is achieved by an introduction of the metal - rich winds into the model with the d / h feature unchanged ( right panels ) . for reference , in both models , the results calculated with a constant sfr coefficient are indicated by dashed lines .
as long as we stick to the conventional scheme of chemical evolution , recent observational implications of ( i ) the present - day high d / h abundance and ( ii ) little evolution of its abundance during the last several gyr in the local disk are hard to believe .
accordingly , the discussion on it so far has been inclined to favor a much lower present - day d / h abundance in terms of an uncertainty in its estimate mainly due to dust correction . here
we present a new view on the d / h evolution in the local disk , based on the theoretical scheme that considers the wind enrichment and the decreasing sfr with a rate of @xmath1 -0.8 gyr@xmath24 , two phenomena that are implied to have occurred in the last several gyr from the observed feature of local disk stars .
the essence of its theory is that all heavy elements are not locally produced but some fraction comes form the outside , thus the models whose mission is the construction of the metal - rich system accept the observed low activity of star formation in the recent gyrs , and end up with the ism rich in d. we insist that the local star formation does not suffice to produce all elements in - situ , and thus d does not destroy so much as expected by the chemical features of the local disk .
similar claim may be seen in the mechanism of radial mixing @xcite , which brings metal - rich stars from the outside ( i.e. , the inner disk ) instead of heavy elements .
@xcite show the star formation history for the local disk which exponentially decreases 1 gyr after the beginning of star formation , as implied by the new reduction of hipparcos data @xcite .
their model succeeds in a high - z stellar abundance system as an outcome of radial migration of stars while it will surely predict a low d today .
we strongly predict an ongoing infall onto the galaxy disk .
our claim is that it operates to reduce the gas abundances in the disk . indeed , today s observed gas abundances are not metal - rich at any places of the disk .
the metallicities of cepheids at the inner disk ( @xmath70 4kpc ) are [ fe / h]@xmath71 at the most @xcite . moreover , a surprisingly rsult is brought by supergiants in two young clusters in the innermost disk at @xmath47=3 - 4 kpc , whose metallicities are found to be [ fe / h]@xmath1 -0.1@xmath49 -0.2 @xcite .
an infall must rain down on today s disk .
discussions with j. bland - hawthorn are gratefully acknowledged .
the author wishes to thank the referee for all valuable comments that helped improve the paper , and is assisted in part by grant - in - aid for scientific research ( 21540246 ) of the japanese ministry of education , culture , sports , science and technology .
99 allende prieto c. et al .
, 2006 , apj , 636 , 804 andrievsky s. m. , luck r. e. , martin p. , lpine j. r. d. , 2004 , a&a , 413 , 159 athanassoula e. , martinez - valpuesta i. , 2009 , in contopoulos c. , patsis p. a. , eds , chaos in astronomy .
springer - verlag , berlin , p. 77
aumer m. , binney , j. j. , 2009 , mnras , 397 , 1286 ballero s. k. , matteucci f. , origlia l. , rich , r. m. , 2007 , a&a , 467 , 123 bekki k. , tsujimoto t. , chiba m. , 2009 , apj , 692 , l24 bensby t. , feltzing s. , lundstrm i. , ilyiin , i. , 2005 , a&a , 433 , 185 bensby t. , zenn a. r. , oey m. s. , feltzing s. , 2007 , apj , 663 , l13 bensby t. et al . , 2010 , a&a , 512 , a41 bland - hawthorn j. , cohen m. , 2003 , apj , 582 , 246 boissier s. , prantzos n. , 1999 , mnras , 307 , 857 bregman j. n. , 1980 , apj , 236 , 577 caimmi r. , 2008 , new astronomy , 13 , 314 ceverino d. , klypin a. , 2009 , apj , 695 , 292 chen l. , hou j. l. , wang j. j. , 2003 , aj , 125 , 1397 chen y. q. et al . , 2008 , aj , 135 , 618 chiappini c. , matteucci f. , romano d. , 2001 , apj , 554 , 1044 chiba m. , beers t. c. , 2000 , aj , 119 , 2843 cignoni m. , degl innocenti s. , prada moroni p. g. , shore s. n. , 2006 , a&a , 459 , 783 clarkson w. et al .
, 2008 , apj , 684 , 1110 clayton d. d. , 1997 , apj , 484 , l67 coc a. , vangioni - flan e. , descouvemont p. , adahchour a. , angulo c. 2004 , apj , 600 , 544 cyburt r. h. , fields b. d. , olive k. a. 2003 , phys .
b , 567 , 227 dame t. m. , 1993 , in holt s. s. , verter f. eds , back to the galaxy .
aip , new york , p. 267
de la fuente marcos r. , de la fuente marcos c. , 2004 , new astron . , 9 , 475 del peloso e. f. , da silva l. , porto de mello g. f. , arany - prado l. i. , 2005 , a&a , 440 , 1153 davies b. , 2009 , apj , 696 , 2014 dunkley j. et al . , 2009 , apjs , 180 , 306 edvardsson b. , andersen j. , gustaffson b. , lambert d. l. , nissen p. e. , tomkin , j. 1993 , a&a , 275 , 101 ellison s. l. , prochaska j. x. , lopez s. 2007 , mnras , 380 , 1245 everett j. e. , schiller q. g. , zweibel e. g. , 2010 , apj , 711 , 13 feltzing s. , gustafsson b. , 1998 , a&as , 129 , 237 feltzing s. , gilmore g. , 2000 , a&a , 355 , 949 flynn c. , holmberg j. , portinari l. , fuchs b. , jahrei@xmath72 h. , 2006 , mnras , 373 , 993 fox a. j. , savage b. d. , wakker b. p. , richter p. , sembach k. r. , tripp t. m. , 2004 , apj , 602 , 738 fox a. j. , wakker b. p. , savage b. d. , tripp t. m. , sembach k. r. , bland - hawthorn j. , 2005 , apj , 630 , 332 fuchs b. , jahrei@xmath72 h. , flynn c. 2009 , aj , 137 , 266 fulbright j. p. , mcwilliam a. , rich r. m. , 2007 , apj , 661 , 1152 geiss j. , gloeckler g. , 1998 , space sci . rev . , 84 , 239 gilmore g. , wyse r. f. g. , jones j. b. , 1995 , aj , 109 , 1095 hbrard g. et al . , 2005 , apj , 65 , 1136 hbrard g. 2010 , in charbonnel c. et al .
, eds , proc .
268 , light elements in the universe .
cambridge univ . press , cambridge ,
hansen b. m. s. et al . , 2002 ,
apj , 574 , l155 hernandez x. , valss - gabaud d. , gilmore g. , 2000 , mnras , 316 , 605 holtzman j. a. et al . , 1993 , aj , 106 , 1826 hou j. l. , prantzos n. , boissier s. , 2000 , a&a , 362 , 921 keeney b. a. , danforth c. w. , stocke j. t. , penton s. v. , shull j. m. , sembach k. r. , 2006 , apj , 646 , 951 kennicutt r. c. , 1998 , arr&a , 36 , 189 kobayashi c. , nomoto k. 2009 , apj , 707 , 1466 kuijken k. , rich r. m. , 2002 , aj , 124 , 2054 kulkarni s. r. , heiles c. , 1987 , in hollenbach d. , thronson h. , eds , interstellar processes .
kluwer , dordrecht , p. 87
launhardt r. , zylka r. , mezger p. g. , 2002 , a&a , 384 , 112 lellouch e. , bzard b. , fouchet t. , feuchtgruber h. , encrenaz t. , de graauw t. , 2001 , a&a , 670 , 610 linsky j. l. et al .
, 2006 , apj , 647 , 1106 lockman f. j. , benjamin r. a. , heroux a. j. , langston g. i. , 2008 , apj , 679 , l21 maciel w. j. , lago l. g. , costa r. d. d. , 2006 , a&a , 453 , 587 matteucci f. , brocato e. , 1990 , apj , 365 , 539 matteucci f. , romano d. , molaro p. , 1999
, a&a , 341 , 458 mcwilliam a. , matteucci f. , ballero s. , rich r. m. , fulbright j. p. , cescutti g. , 2008 , aj , 136 , 367 melndez j. et al . , 2008 , a&a , 484 , l21 murray n. , quataert e. , thompson , t .
a. , 2005 , apj , 618 , 569 nordstrm b. et al . , 2004 ,
a&a , 418 , 989 omeara j. m. , burles s. , prochaska j. x. , prochter g. e. , bernstein r. a. , burgess k. m. , 2006 , apj , 649 , l61 onodera s. et al . , 2010 , submitted to apjl ( arxiv:1006.5764 ) oppenheimer b. d. , dav r. , 2006 , mnras , 373 , 1265 oppenheimer b. d. et al . , 2010 , mnras , in press ortolani s. , renzini a. , gilmozzi r. , marconi g. , barbuy b. , bica e. , rich r. m. , 2005 , nature , 377 , 701 pagel b. e. j. , 1997 , nucleosynthesis and chemical evolution of galaxies , cambridge univ . press , cambridge peek j. e. g. , puman m. e. , sommer - larsen j. , 2008 , apj , 674 , 227 pettini m. , zych b. j. , murphy m. t. , lewis a. , steidel c. c. , 2008 , mnras , 391 , 1499 prantzos n. , 1996 , a&a , 310 , 106 prodanovi t. , fields b. d. , 2008 , jcap , 9 , 3 prodanovi t. , steigman g. , fields b. d. , 2010 , mnras , 406 , 1108 reddy b. e. , tomkin j. , lambert d. l. , allende prieto c. , 2003 , mnras , 340 , 304 rocha - pinto h. j. , maciel w. j. , 1996 , mnras , 279 , 447 rocha - pinto h. j. , scalo j. , maciel w. j. , flynn c. , 2000 , a&a , 358 , 869 romano d. , tosi m. , chiappini c. , matteucci f. 2006 , mnras , 369 , 295 rokar r. , debattista v. p. , stinson g. s. , quinn t. r. , kaufmann t. , wadsley j. , 2008a , apj , 675 , l65 rokar r. , debattista v. p. , quinn t. r. , stinson g. s. , wadsley j. 2008b , apj , 684 , l79 roychowdhury s. , chengalur j. n. , begum a. , karachentsev d. , 2009 , mnras , 397 , 1435 snchez - blzquez p. , courty s. , gibson b. k. , brook c. b. , 2009 , mnras , 398 , 591 savage b. d. , lehner n. , fox a. , wakker b. , sembach k. , 2007 , apj , 659 , 1222 schenrich r. , binney j. , 2009 , mnras , 396 , 203 sellwood j. a. , binney j. j. , 2002 , mnras , 336 , 785 sembach k. r. et al . , 2003 , apjs , 146 , 165 sembach k. r. , 2004 , apjs , 150 , 387 shapiro p. r. , fields g. b. 1976 , apj , 205 , 762 stanghellini l. , haywood m. , 2010 , apj , 714 , 1096 steigman g. , romano d. , tosi m. 2007 , mnras , 378 , 576 thom c. , peek j. e. g. , putman m. e. , heiles c. , peek k. m. g. , wilhelm r. , 2008 , apj , 684 , 364 tosi m. , 2010 , in charbonnel c. et al .
, eds , proc . iau symp .
268 , light elements in the universe .
cambridge univ .
press , cambridge , p. 153 tosi m. , steigman g. , matteucci f. , chiappini c. , 1998 , apj , 498 , 226 tsujimoto t. , nomoto k. , yoshii y. , hashimoto m. , yanagida s. , thielemann f .- k . , 1995 , mnras , 277 , 945 tsujimoto t. , yoshii y. , nomoto k. , matteucci f. , thielemann f .- k .
, & hashimoto m. , 1997 , apj , 483 , 228 tsujimoto t. , 2007 , apj , 665 , l115 tsujimoto t. , bland - hawthorn j. , freeman k. c. , 2010 , pasj , 62 , 447 vergely j .-
, kppen j. , egret d. , bienaym o. , 2002 , a&a , 390 , 917 yin j. et al . , 2009 , a&a , 505 , 497 yoshii y. , tsujimoto t. , nomoto k. , 1996 , apj , 462 , 266 wakker b. p. et al . , 1999 , nature , 402 , 388 wolfe a. m. , gawiser e. , prochaska j. x. , 2005 , ara&a , 43 , 861 wyder t. k. et al .
, 2009 , apj , 696 , 1834 wyse r. f. g. , gilmore g. , 1995 , aj , 110 , 2771 zoccali m. et al . , 2003 ,
a&a , 399 , 931 | chemical features of the local stellar disk have firmly established that long - term , continuous star formation has been accompanied by a steady rate of accretion of low - metallicity gas from the halo .
we now argue that the recent discovery of an enhanced deuterium ( d ) fraction in the galaxy is consistent with this picture .
we consider two processes : the destruction of d in the interior of stars ( astration ) and the supply of nearly primordial d associated with the gas infall .
conventional galactic chemical evolution models predict a monotonic decrease in d / h with time with a present - day d / h abundance which is much lower than the local value recently revealed .
this predicted feature is the result of high levels of deuterium astration involved in the formation of the local metal - enhanced disk .
here we propose a new channel to explain the observed enhancement in d / h . our model , which invokes ongoing gaseous infall and a star formation rate that declines over the past several gyr , predicts that the d astration is suppressed over the same time interval .
[ firstpage ] ism : abundances galaxy : disc galaxy : evolution stars : abundances . |
nowadays it is commonly believed that neutrinos are massive particles and there is a nonzero mixing between different neutrino generations .
these neutrino properties result in transitions between neutrino flavors , or neutrino oscillations .
it is also known that various external fields , like electroweak interaction of neutrinos with background fermions , neutrino electromagnetic interaction , and neutrino interaction with a strong gravitational field , can also influence the process of neutrino oscillations . as shown in ref . , noninertial effects in accelerated and rotating frames can affect neutrino propagation and oscillations .
the consideration of the reference frame rotation is particularly important for astrophysical neutrinos emitted by a rapidly rotating compact star , e.g. , a pulsar .
for example , the possibility of the pulsar spin down by the neutrino emission and interaction with rotating matter was recently discussed in ref . .
it should be noted that besides elementary particle physics , various processes in noninertial frames are actively studied in condensed matter physics .
for instance , the enhancement of the spin current in a semiconductor moving with an acceleration was recently predicted in ref . .
in the present work we summarize the results of ref . , where the neutrino interaction with background matter in a rotating frame was studied .
we assume that neutrinos can interact with background fermions by means of electroweak forces .
we also take that neutrino mass eigenstates are dirac particles . in our treatment
we account for the noninertial effects since we find the exact solution of a dirac equation for a neutrino moving in the curved space - time with a metric corresponding a rotating frame .
this work is organized in the following way . in sec .
[ sec : numatfls ] , we start with the brief description of the neutrino interaction with background matter in minkowski space - time .
we consider matter moving with a constant velocity and discuss both neutrino flavor and mass eigenstates .
then , in sec . [ sec : massnunonif ] , we study background matter moving with an acceleration .
the dirac equation , including noninertial effects , for a neutrino interacting with such matter is written down in a frame where matter is at rest . in sec .
[ sub : rotation ] , we solve the dirac equation and find the neutrino energy spectrum for ultrarelativistic neutrinos moving in matter rotating with a constant angular velocity .
then , in sec . [ sec : osc ] , we apply our results for the description of neutrino oscillations in rotating background matter . the effective schrdinger equation governing neutrino oscillations is derived and the new resonance condition is obtained .
we consider how the matter rotation can affect the resonance in neutrino oscillation in a realistic astrophysical situation . finally , in sec .
[ sec : summary ] , we summarize our results .
in this section we describe the interaction of different neutrino flavors with background matter in a flat space - time .
we discuss a general case of matter moving with a constant mean velocity and having a mean polarization .
then we consider the matter interaction of neutrino mass eigenstates , which are supposed to be dirac particles .
the interaction of the neutrino flavor eigenstates @xmath0 , @xmath1 , with background matter in the flat space - time is described by the following effective lagrangian@xcite : @xmath2 where @xmath3 , @xmath4 are the dirac matrices , and @xmath5 . the interaction lagrangian in eq .
( [ eq : lf ] ) is derived in the mean field approximation using the effective external currents @xmath6 depending on the characteristics of background matter as@xcite @xmath7 where @xmath8 is the fermi constant and the sum is taken over all background fermions @xmath9 .
here @xmath10 is the hydrodynamic current and @xmath11 is the four vector of the matter polarization . in eqs .
( [ eq : jms ] ) and ( [ eq : lambdams ] ) , @xmath12 is the invariant number density ( the density in the rest frame of fermions ) , @xmath13 is the invariant polarization ( the polarization in the rest frame of fermions ) , and @xmath14 is the four velocity . to derive eqs .
( [ eq : flavnueffp])-([eq : lambdams ] ) it is crucial that background fermions have constant velocity . only in this situation
one can make a boost to the rest frame of the fermions where @xmath12 and @xmath13 are defined .
the explicit form of the coefficients @xmath15 in eq .
( [ eq : flavnueffp ] ) can be found in ref . .
nowadays it is experimentally confirmed that the flavor neutrino eigenstates are the superposition of the neutrino mass eigenstates , @xmath16 , @xmath17 , @xmath18 where @xmath19 is the unitary mixing matrix . the transformation in eq .
( [ eq : nupsi ] ) diagonalizes the neutrino mass matrix . only using the neutrino mass eigenstates we can reveal the nature of neutrinos
, i.e. say whether they are dirac or majorana particles . despite the great experimental efforts to shed light upon the nature of neutrinos ,
this issue still remains open .
here we shall suppose that @xmath16 correspond to dirac fields .
the effective lagrangian for the interaction of @xmath16 with background matter can be obtained using eqs .
( [ eq : lf ] ) and ( [ eq : nupsi ] ) , @xmath20 where @xmath21 is the nondiagonal effective potential in the mass eigenstates basis . using eq .
( [ eq : lg ] ) one obtains that the corresponding dirac equations for the neutrino mass eigenstates are coupled , @xmath22 \psi_{i } = \sum_{j\neq i}\gamma_{\mu}^{\mathrm{l}}g_{ij}^{\mu}\psi_{j},\ ] ] where @xmath23 is the mass of @xmath16 .
one can proceed in the analytical analysis of eq .
( [ eq : depsi ] ) if we exactly account for only the diagonal effective potentials @xmath24 . to take into account the r.h.s . of eq .
( [ eq : depsi ] ) , depending on the nondiagonal elements of the matrix @xmath25 , with @xmath26 , one should apply a perturbative method ( see sec .
[ sec : osc ] below ) .
in this section we generalize the dirac equation for a neutrino interacting with a background matter to the situation when the velocity of the matter motion is not constant .
in particular , we study the case of the matter rotation with a constant angular velocity .
then we obtain the solution of the dirac equation and find the energy spectrum .
if we discuss a neutrino mass eigenstate propagating in a nonuniformly moving matter , the expressions for @xmath6 in eqs .
( [ eq : flavnueffp])-([eq : lambdams ] ) become invalid since they are derived under the assumption of the unbroken lorentz invariance .
the most straightforward way to describe the neutrino evolution in matter moving with an acceleration is to rewrite the dirac equation for a neutrino in the noninertial frame where matter is at rest . in this case
one can unambiguously define the components of @xmath6 .
assuming that background fermions are unpolarized , we find that in this reference frame @xmath27 with the rest of the effective potentials being equal to zero .
it is known that the motion of a test particle in a noninertial frame is equivalent to the interaction of this particle with a gravitational field . the dirac equation for a massive neutrino moving in a curved space - time and interacting with background matter
can be obtained by the generalization of eq .
( [ eq : depsi ] ) ( see also ref . ) , @xmath28 \psi = \frac{1}{2}\gamma_{\mu}(x)g^{\mu } \left [ 1-\gamma^{5}(x ) \right ] \psi,\ ] ] where @xmath29 are the coordinate dependent dirac matrices , @xmath30 is the covariant derivative , @xmath31 is the spin connection , @xmath32 , @xmath33 is the covariant antisymmetric tensor in curved space - time , and @xmath34 is the determinant of the metric tensor @xmath35 .
note that in eq .
( [ eq : depsicurv ] ) we account for only the diagonal neutrino interaction with matter .
that is why we omit the index @xmath36 in order not to encumber the notation : @xmath37 etc .
it should be noted that analogous dirac equation was discussed in ref . .
we shall be interested in the neutrino motion in matter rotating with the constant angular velocity @xmath38 .
choosing the corotating frame we get that only @xmath39 is nonvanishing , cf .
( [ eq : gab ] ) and ( [ eq : f0nonin ] ) .
the interval in the rotating frame is@xcite @xmath40 where we use the cylindrical coordinates @xmath41 .
one can check that the metric tensor in eq .
( [ eq : mertrot ] ) can be diagonalized , @xmath42 , if we use the following vierbein vectors : @xmath43 here @xmath44 is the metric in a locally minkowskian frame .
let us introduce the dirac matrices in a locally minkowskian frame by @xmath45 , where @xmath46 is the inverse vierbein : @xmath47 .
starting from now , we shall mark an index with a bar to demonstrate that a gamma matrix is defined in a locally minkowskian frame . as shown in ref .
, @xmath48 does not depend on coordinates .
after the straightforward calculation of the spin connection on the basis of eq .
( [ eq : vierb ] ) , the dirac eq .
( [ eq : depsicurv ] ) can be rewritten as @xmath49\psi = \frac{1}{2 } \sqrt{1-\omega^{2}r^{2}}\gamma^{\bar{0}}g^{0 } ( 1-\gamma^{\bar{5}})\psi , \nonumber \\
\mathcal{d } = & \mathrm{i } \frac{\gamma^{\bar{0}}+\omega r\gamma^{\bar{2}}}{\sqrt{1-\omega^{2}r^{2 } } } \partial_{0 } + \mathrm{i } \gamma^{\bar{1 } } \left ( \partial_{r}+\frac{1}{2r } \right ) + \mathrm{i } \gamma^{\bar{2 } } \frac{\sqrt{1-\omega^{2}r^{2}}}{r}\partial_{\phi } + \mathrm{i } \gamma^{\bar{3}}\partial_{z } \notag \\ & - \frac{\omega}{2(1-\omega^{2}r^{2 } ) } \gamma^{\bar{3}}\gamma^{\bar{5}}.\end{aligned}\ ] ] the analogous dirac equation was recently derived in ref . .
since eq .
( [ eq : derotf ] ) does not explicitly contain @xmath50 , @xmath51 , and @xmath52 , its solution can be expressed as @xmath53 where @xmath54 is the spinor depending on the radial coordinate , @xmath55 ( see , e.g. , ref . ) , and @xmath56 . in eq .
( [ eq : derotf ] ) one can neglect terms @xmath57 .
indeed , if we study a neutrino in a rotating pulsar , then @xmath58 and @xmath59 .
thus @xmath60 is a small parameter .
therefore eq .
( [ eq : derotf ] ) can be transformed to @xmath61 \psi_{r}=0,\end{gathered}\ ] ] where we keep only the terms linear in @xmath38 .
it should be noted that the term @xmath62 in eq .
( [ eq : depsisimp ] ) is equivalent to the neutrino interaction with matter moving with an effective velocity . the solution of eq .
can be presented in the form@xcite , @xmath63 and @xmath64 , where @xmath65 here @xmath66 , @xmath67 , @xmath68 is the laguerre function , and @xmath69 .
the explicit form of the laguerre function can be found , e.g. , in ref . .
to derive eq .
we use the dirac matrices in the chiral representation@xcite . in the important case
when @xmath70 , the coefficients @xmath71 , @xmath72 , in eq .
are expressed in the following way@xcite : @xmath73 it should be noted that the solutions presented in eqs .
( [ eq : etaxi ] ) and ( [ eq : c1 - 4slowrot ] ) satisfy the normalization condition , @xmath74 here @xmath75 and @xmath76 are related by eq .
( [ eq : psitildepsi ] ) .
the energy levels in eq .
are @xmath77^{2 } = & ( 2n\omega)^{2}+4n\omega g^{0 } + \left ( p_{z}-\frac{\omega}{2 } \right)^{2 } , \nonumber \\ \left [ e_{s}-2n\omega \right]^{2 } = & ( 2n\omega)^{2 } + \left ( p_{z}+\frac{\omega}{2 } \right)^{2},\end{aligned}\ ] ] where @xmath78 and @xmath79 are the energies of active and sterile neutrinos respectively . comparing the expression for @xmath80 with the energy of a neutrino in an inertial nonrotating frame @xmath81 , where @xmath82 is the momentum in the equatorial plane
, we can identify @xmath83 inside the square root as @xmath84 .
it should be also noted that the term @xmath83 , which additively enters to both @xmath78 and @xmath79 , is due to the noninertial effects for a dirac fermion in a rotating frame@xcite .
we can also get the corrections to the energy levels due to the nonzero mass , @xmath85 . on the basis of eq .
( [ eq : etaxi ] ) and ( [ eq : c1 - 4slowrot ] ) one finds the expression for @xmath86 in the limit @xmath70 , @xmath87 if we discuss neutrinos moving along the rotation axis , then @xmath88 . using eq .
( [ eq : energylev ] ) we get the energy levels of active neutrinos in this case @xmath89 where we also keep the mass correction in eq .
( [ eq : eas1 ] ) .
one can see in eq .
( [ eq : eadecom ] ) that @xmath90 corresponds to the energy of a left - handed neutrino interacting with background matter in a flat space - time . the rest of the terms in eq .
( [ eq : eadecom ] ) are the corrections due to the matter rotation .
in this section we study the evolution of the system of massive mixed neutrinos in rotating matter . we formulate the initial condition for this system and
derive the effective schrdinger equation which governs neutrino flavor oscillations .
then we find the correction to the resonance condition owing to the matter rotation and estimate its value for a millisecond pulsar .
we can generalize the results of sec .
[ sec : massnunonif ] to include different neutrino eigenstates .
the interaction of neutrino mass eigenstates with background matter is nondiagonal , cf .
( [ eq : lg ] ) . therefore the generalization of eq .
( [ eq : derotf ] ) for several mass eigenstates @xmath16 reads @xmath91 \psi_{i } = \frac{1}{2}\gamma^{\bar{0 } } g_{i}^{0 } ( 1-\gamma^{\bar{5 } } ) \psi_{i } + \frac{1}{2}\gamma^{\bar{0 } } \sum_{j\neq i } g_{ij}^{0}(1-\gamma^{\bar{5}})\psi_{j},\ ] ] where @xmath92 and @xmath93 are the time components of the matrix @xmath25 given in eq .
( [ eq : gab ] ) , @xmath23 is the mass of @xmath16 , and @xmath94 can be found in eq .
( [ eq : derotf ] ) . as in sec .
[ sec : massnunonif ] , we omitted the term @xmath95 in eq .
( [ eq : derotfpsia ] ) .
note that eq .
( [ eq : derotfpsia ] ) is a generalization of eq .
( [ eq : depsi ] ) for a system of the neutrino mass eigenstates moving in a rotating frame .
we shall study the evolution of active ultrarelativistic neutrinos and neglect neutrino - antineutrino transitions . in this case
we can restrict ourselves to the analysis of two component spinors .
the general solution of eq .
( [ eq : derotfpsia ] ) has the form , @xmath96 where @xmath97 are the basis spinors and @xmath98 are the @xmath99-number functions .
the energy levels @xmath100 are given in eq .
( [ eq : eadecom ] ) with @xmath101 . here
we omit the subscript @xmath102 in order not to encumber the notation .
our goal is to find the coefficient @xmath103 .
we neglect the small ratio @xmath104 in eq .
( [ eq : gensolosc ] ) . considering the system of two neutrino mass eigenstates , @xmath105 , parameterized with one mixing angle @xmath106 , and choosing the appropriate initial condition@xcite , on the basis of eq .
( [ eq : derotfpsia ] ) we get the effective schrdinger equation for @xmath107,@xmath108 \\
g_{12}^{0 } \exp \left [ \mathrm{i } \left ( e_{2}-e_{1 } \right ) t \right ] & 0 \end{array } \right ) \tilde{\psi}.\ ] ] here we omitted all the indexes of @xmath109 besides @xmath105 .
it is convenient to introduce the modified effective wave function @xmath110 , where @xmath111 , @xmath112 . using eq .
( [ eq : scheqtildepsi ] ) , we get for @xmath113 @xmath114 note that eq .
( [ eq : scheqpsi ] ) has the form of the effective schrdinger equation one typically deals with in the study of neutrino flavor oscillations in background matter .
if the transition probability for @xmath115 is close to one , i.e. @xmath116 , flavor oscillations of neutrinos are said to be at resonance .
using eqs .
( [ eq : nupsi ] ) , ( [ eq : f0nonin ] ) , ( [ eq : eadecom ] ) , and , the resonance condition can be written as , @xmath117 where @xmath118 is the mass squared difference .
let us consider electroneutral background matter composed of electrons , protons , and neutrons .
if we study the @xmath119 oscillation channel , where @xmath120 , we get that @xmath121 and @xmath122 , where @xmath123 and @xmath124 are the densities of electrons and neutrons . using eq .
( [ eq : rescondgen ] ) , we obtain that @xmath125 at the absence of rotation , @xmath126 , eq .
( [ eq : resnuenumu ] ) is equivalent to the mikheyev - smirnov - wolfenstein resonance condition in background matter@xcite .
let us evaluate the contribution of the matter rotation to the resonance condition in eq .
( [ eq : resnuenumu ] ) for a neutrino emitted inside a rotating pulsar .
we make a natural assumption that for a corotating observer neutrinos are emitted in a spherically symmetric way from a neutrinosphere .
that is we should take that @xmath127 and @xmath128 . then the trajectory of a neutrino is deflected because of the noninertial effects and the interaction with background matter .
the radius @xmath129 of the trajectory can be found from @xmath130 where we take into account that @xmath131 .
we shall assume that @xmath132 , where @xmath133 is the pulsar radius .
in this case neutrinos escape a pulsar .
taking that @xmath134 and using eq .
( [ eq : trajrad ] ) , we get that the correction to the resonance condition in eq .
( [ eq : resnuenumu ] ) is @xmath135 .
the obtained correction to the effective number density is small but nonzero .
this result corrects our previous statement@xcite that a matter rotation does not contribute neutrino flavor oscillations .
in conclusion we notice that we have studied the evolution of massive mixed neutrinos in nonuniformly moving background matter .
the interaction of neutrinos with background fermions is described in frames of the fermi theory ( see sec .
[ sec : numatfls ] ) . a particular case of the matter rotating with a constant angular velocity has been studied in sec .
[ sub : rotation ] .
we have derived the dirac equation for a weakly interacting neutrino in a rotating frame and found its solution in case of ultrarelativistic neutrinos , cf .
( [ eq : etaxi ] ) and ( [ eq : c1 - 4slowrot ] ) . the energy spectrum obtained in eqs .
( [ eq : energylev ] ) and ( [ eq : eas1 ] ) includes the correction owing to the nonzero neutrino mass .
we have used the dirac equation in a noninertial frame , cf .
( [ eq : depsicurv ] ) , as a main tool for the study of the neutrino motion in matter moving with an acceleration . to develop the quantum mechanical description of such a neutrino we have chosen a noninertial frame where matter is at rest . in this frame
the effective potential of the neutrino - matter interaction is well defined .
however , the wave equation for a neutrino turns out to be more complicated since one has to deal with noninertial effects . in sec .
[ sec : osc ] we have generalized our results to include various neutrino generations as well as mixing between them .
we have derived the effective schrdinger equation which governs neutrino flavor oscillations .
we have obtained the correction to the resonance condition in background matter owing to the matter rotation .
studying neutrino oscillations in a millisecond pulsar , we have obtained that the effective number density changes by @xmath136 owing to the matter rotation . despite the obtained correction
is small , we may suggest that our results can have some implication to the explanation of great linear velocities of pulsars .
it was suggested in ref . that an asymmetry in neutrino oscillations in a magnetized pulsar can explain a great linear velocity of the compact star .
an evidence for the alignment of the angular and the linear velocity vectors of pulsars was reported in ref . .
therefore we may suggest that neutrino flavor oscillations in a rapidly rotating pulsar can contribute to its linear velocity .
it should be noted that neutrino spin - flavor oscillations , including noninertial effects , in a rapidly rotating magnetized star were studied in ref . in the context of the explanation of high linear velocities of pulsars . finally , we mention that the dirac equation for a fermion , electroweakly interacting with the rotating background matter , was recently solved@xcite .
the vierbein vectors , different from these in eq . , were used in ref . .
comparing the energy levels obtained in ref . with the results of the general analysis@xcite ,
one concludes that the vierbein used in ref .
is more appropriate for the description of ultrarelativistic particles like neutrinos .
i am thankful to the organizers of the international conference on massive neutrinos for the invitation and support , to s. p. gavrilov for helpful comments , to fapesp ( brazil ) for the grant no .
2011/50309 - 2 , to the tomsk state university competitiveness improvement program and to rfbr ( research project no . 15 - 02 - 00293 ) for partial support .
g. lambiase , neutrino oscillations in non - inertial frames and the violation of the equivalence principle .
neutrino mixing induced by the equivalence principle violation , _ eur .
phys . j. c _ * 19 * , 553 ( 2001 ) .
s. johnston , g. hobbs , s. vigeland , m. kramer , j. m. weisberg and a. g. lyne , evidence for alignment of the rotation and velocity vectors in pulsars , _ mon . not .
soc . _ * 364 * , 1397 ( 2005 ) [ astro - ph/0510260 ] . | we study dirac neutrinos propagating in rotating background matter .
first we derive the dirac equation for a single massive neutrino in the noninertial frame , where matter is at rest .
this equation is written in the effective curved space - time corresponding to the corotating frame .
we find the exact solution of the dirac equation .
the neutrino energy levels for ultrarelativistic particles are obtained .
then we discuss several neutrino mass eigenstates , with a nonzero mixing between them , interacting with rotating background matter .
we derive the effective schrdinger equation governing neutrino flavor oscillations in rotating matter .
the new resonance condition for neutrino oscillations is obtained .
we also examine the correction to the resonance condition caused by the matter rotation . |
the launch of the rossi x - ray timing explorer ( rxte ) in december 1995 marked a new beginning for the study of compact objects and their accretion disks . since then quasi - periodic oscillations ( qpos ) in the kilo - hertz range ( khz qpos henceforth ) have been discovered in the persistent flux of a number of low mass x - ray binaries ( lmxbs ) , including both the z and the atoll sources , as well as sources that have not been classified . for a comprehensive review of the literature on this subject , see van der klis ( 1997 ) .
although questions about the detailed physical mechanism for the generation of these oscillations has not been settled , it is clear that the khz qpos are related to the dynamic time scale near the neutron star surface , and the ncos likely to be the neutron star spin .
one of the puzzling characteristics of the khz qpos is their similarity across the large number of sources which cover two decades in luminosity and probably also two decades in magnetic field .
this characteristic raises a serious question as to what are the roles that the overall accretion rate and the neutron star magnetic field play in determining the qpo frequencies and other characteristics .
it places a severe constraint on any potential model for their generation mechanism .
another puzzling characteristic of both the khz qpos and the ncos is that the inferred neutron star spin frequencies of these lmxbs are in a very narrow range in the vicinity of 300 hz ( @xcite ) .
both of these puzzles are hard to understand in the context of the standard disk magneto - dynamic interactions ( @xcite ) without invoking some kind of correlation between the magnetic field and the mass accretion rate across these sources ( @xcite ) , unless we accept the possibility that either the inferred frequencies from the burst oscillations are not directly related to the neutron spins or the disk magneto - dynamic interactions picture we have at the present is significantly incomplete .
in this paper we report rxte observations of aql x-1 during its february 1997 outburst .
we present analysis results of the khz qpos in the persistent flux and the ncos during a type - i burst .
we believe that the data strongly indicate that there exist new regimes characterized by different branches in the relationship between the khz qpo frequency and the total source luminosity .
the data analyzed for this paper results from 12 separate rxte observations or pointings at aql x-1 covering the calendar period from 16 february through 10 march 1997 .
each observation is typically two rxte orbits in duration and results in about 8,000 s of good data segmented into two to three pieces due to rxte passages through the south atlantic anomaly and earth occult of aql x-1 .
the typical time gap between two consecutive observations is two days .
the first observation took place approximately a week after aql x-1 passed the peak luminosity and the last when the source had essentially reached its quiescent state .
two type - i x - ray bursts occurred during these observations , but only one of which has data with good time and energy spectral resolution .
the other one occurred shortly after an earth occult before the instrument was completely ready for taking data .
in addition to the standard 1 and 2 data sets , we have event mode data with 64 energy channels and 128@xmath2 ( @xmath2 = @xmath3 ) time resolution , and a burst catcher mode with 1 energy channel and 128@xmath2 time resolution . during the type - i burst the event mode data exceeds the bandwidth between the pca and the spacecraft , therefore has gaps .
we have used the burst catcher data to perform the timing analysis reported below .
the upper panel of figure [ pca_lc_and_rms ] shows the light curve from the entire data set .
each point represents data from one rxte orbit .
the 12 intervals bracketed by the vertical lines represent the 12 pointings .
we have fitted the energy spectrum with a model consisting of a blackbody component and a power - law component with an absorption column density fixed at @xmath4 @xmath5 ( @xcite ) .
hereafter , flux values are calculated using the fitted model and parameters .
we have constructed average fft power spectrum for data from every orbit of the 12 pointings using event mode data .
only 2 pointings have shown clearly detectable khz qpo peaks in their power spectra , the 7th and the 8th which took place on 27 february and 1 march , respectively .
the lower panel of figure [ pca_lc_and_rms ] shows the corresponding upper limits for those days when khz qpos are not positively detected . the overall average fractional rms amplitude in the rxte / pca band is ( [email protected])% .
the width of the qpo peak varies with time , but does not correlate with any other variable .
it has an overall average of 6.23 hz ( fwhm ) ( see figure [ aqlx1_khz_lineup ] ) .
figure [ freq_vs_flux ] shows the correlation between the qpo centroid frequency and the measured flux in the 2 - 10 kev band .
the correlation between the frequency and count rate is very similar except that there appears to be more scatter among the points .
in particular we have also examined the correlation between the qpo frequency and the flux in the blackbody component and found that the correlation is much worse .
this is in contrast to the case of 4u 0614 + 091 ( @xcite ) where the correlation between the two seem to be rather good over a long time scale .
we note that on each day individually , the qpo frequency appears to be correlated with the flux , but there is a shift between the two days .
the 7 points toward the lower left part of the graph which stand away from the rest of the points are there because there was a 10% drop in overall flux right after the type - i burst .
this kind of correlation on short time scale , i.e. , from minutes to hours , and the apparent lack of it on longer time scale , i.e. , days , between the qpo frequency and the overall flux or count rate has been observed before in other sources , e.g. , in 4u 1608 - 52 ( @xcite ) and 4u 0614 + 091 ( @xcite ) .
it showed up very dramatically in these observations of aql x-1 .
the rms amplitude varies with energy .
they are @xmath7% and @xmath8% in the 2 - 5 kev and 5 - 10 kev bands , respectively .
since only one qpo peak is observed on each of the two days , it is possible that , even though they cover a similar range in frequency , they could be the manifestations of the two qpo peaks that have been commonly observed in other sources , e.g. , 4u 1728 - 34 . to investigate this possibility ,
we have compared the dynamic fft power spectrum of 4u 1728 - 34 obtained over a two day period in february 1996 and that of aql x-1 on 27 february and 1 march 1997 .
we have found that in the case of 4u 1728 - 34 , the qpo with the lower frequency is always much narrower in width than the qpo with the higher frequency .
they can be distinguished on power spectral width alone . on the other hand ,
the qpo peaks of aql x-1 during the two days look indistinguishable in width . in particular
, their widths are very similar to the widths of the lower frequency qpo peak of 4u 1728 - 34 .
therefore we conclude that most likely the aql x-1 qpos we have observed during the two days are the same qpo peak . following mendez _
et al._(1997 ) we have also searched for a second qpo peak by fitting the power spectrum from every 64 s of data and using the frequency from the fit to align all the power spectra . as shown in figure [ aqlx1_khz_lineup ] we have found no indication of a second peak anywhere in the resulting power spectrum .
in particular we have examined the part of the power spectrum where one would expect a second peak , namely 275 hz above or below the existing qpo peak and found no statistically significant enhancement at all .
the 95% confidence level upper limit is 1% .
an fft of the 14 s time series of the burst catcher data reveals that its intensity oscillates at frequencies near 549 hz , as shown by the dashed histograms in figure [ burst_psd ] . to investigate the frequency and its amplitude variation with time , we have constructed power spectra for each of the 14 seconds of data , as shown in figure [ burst_rms_vs_time ] .
it can be seen that statistically speaking , the oscillations do not start until after the peak of the light curve which is shown in the lower panel . to investigate whether the oscillation during the burst is coherent ,
we have performed a search in @xmath9 space for optimal values of @xmath10 and @xmath1 to minimize the width of the fft peak .
the detailed procedure is that , given a trial pair of ( @xmath11 ) , we first transform the original arrival time @xmath12 of each photon to @xmath13 and then construct the fft power spectrum using the new time series of @xmath14 . using the 8 s of data right after the peak of the light curve , we have found that when @xmath15 hz and @xmath1=0.14 hz s@xmath16 , we obtain the power spectrum with the minimal width , as shown by the solid histogram in figure [ burst_psd ] .
the amount of power in the peaks of the dashed and solid histograms is @xmath17 and @xmath18 , respectively . in other words , we have gathered all the powers from a much larger frequency region into a much smaller region by taking into account a @xmath1 .
the width of the resultant peak is such that , though it is not consistent with perfect coherence , it is fairly close .
therefore we conclude that this is strong evidence to support the interpretation that the observed oscillations are due to the spin of the neutron star photosphere .
the large @xmath1 indicates that the photosphere is probably mostly decoupled from the rest of the star and therefore conserves its own angular momentum in the course of contraction ( @xcite ) .
the inclusion of a @xmath19 does not decrease the width of the peak .
since the energy spectrum of the source changed very little , if any , from 27 february to 1 march , it is quite reasonable to assume that the measured flux and the true mass accretion rate are positively correlated .
figure [ freq_vs_flux ] clearly indicates that it is not possible that , given a mass accretion rate , or for that matter , given an x - ray flux level , the qpo frequency is uniquely determined . on the contrary
, figure [ freq_vs_flux ] clearly indicates that at least two flux levels can give the same qpo frequency .
we believe this is a clear manifestation in a single source of the puzzle we discussed in the introduction , i.e. , on the one hand , there appears to be a correlation between the qpo frequency and the source luminosity for individual sources ; on the other hand , despite the vast difference in mass accretion rates , the qpo frequencies are essentially the same among all those sources .
it is therefore reasonable to conclude that , in lmxbs , regimes with essentially identical qpo frequency and energy spectrum can exist at very different fluxes or mass accretion rates .
this conclusion is inevitable when one further examines figure [ freq_vs_flux ] .
for example , on 27 february 1997 , as the flux decreased , the qpo frequency also decreased .
had the observation been carried out continuously from 27 feb through 1 march , we would have had to have observed one of the following three possibilities : ( 1 ) at some point the khz qpos cease being observable for some time and then becoming observable again at a higher frequency ; ( 2 ) at first the qpo frequency is positively correlated with the flux until the frequency reaches a smaller value than 750 hz , then they become anti - correlated for some time so that as the flux decreases , the qpo frequency increases until it reaches above 830 hz and the anti - correlation changes back to the positive correlation ; ( 3 ) as the flux decrease , there is a sudden and discontinuous jump in the qpo frequency from below 750 hz to above 830 hz . in all likelihood ,
possibility ( 1 ) probably would have been observed .
the reason is that , to our knowledge , the anti - correlation required by possibility ( 2 ) has never been observed in any source at any time .
neither has the sudden and discontinuous change in qpo frequency of possibility ( 3 ) .
the data we have presented have naturally led us to the conclusion that the total x - ray flux , therefore the total mass accretion rate , of a source can not uniquely determine the frequency of the khz qpos .
moreover we infer from the data that there exist many similar regimes for the inner region of the accretion disk . in each of these regimes ,
the khz qpo frequency is more or less tightly correlated with the mass accretion rate or the total x - ray flux .
the qpo frequency ranges of these regimes are rather similar .
they are all in the vicinity of 1,000 hz .
the disk can transition from one regime to another with change in the overall mass accretion rate .
this transition of a lmxb from one state to another with change in mass accretion rate is not new .
for example , the z source gx 5 - 1 has been believed to move along its z track on the x - ray color - color diagram as the mass accretion changes ( @xcite ) .
this motion along the z track , however , involves significant changes in the energy spectral shape .
what is new with aql x-1 is that , even though the mass accretion rate and the energy spectral state can be different from time to time for a given source , the geometric location ( as inferred from the khz qpo frequency ) of the disk with respect to the neutron star does not seem to change by much .
this is just another way of saying that the mass accretion rate does not directly determine the qpo frequency .
it is the disk regime that determines the large scale of the khz qpo frequency and the mass accretion rate affects it in the smaller scale in between regime changes .
we concluding by noting that a possible scenario explaining the observed two regimes on the qpo frequency vs. flux plane is that the disk structure changed during the decay of the outburst .
the different regimes correspond to different vertical scale heights .
therefore khz qpos can be used to probe the accretion disk structure evolution .
berger , m. , van der klis , m. , van paradijs , j. , lewin , w. h. g. , lamb , f. , vaughan , b. , kuulkers , e. , augusteijn , t. , zhang , w. , marshall , f. e. , swank , j. h. , lapidus , i. , lochner , j. c. , strohmayer , t. e. 1996 , , 469 , l13 | the rossi x - ray timing explorer observed the soft x - ray transient aql x-1 during its outburst in february and march 1997 .
we report the discovery of quasi - periodic oscillations ( qpos ) in its persistent flux with frequencies in the range of 740 to 830 hz , q - value of over 100 , and a fractional rms amplitude of @xmath0% , and nearly coherent oscillations ( ncos ) during a type - i burst with a frequency of 549 hz .
the frequency of the qpos in the persistent flux is correlated with the mass accretion rate on time scale of hours , but not on time scale of days .
this is most likely the manifestation in a single source of the khz qpo puzzle observed among many sources , i.e. , on the one hand , individual sources show a correlation between the qpo frequency and the inferred mass accretion rate , on the other hand , the dozen or so sources with luminosities spanning two decades have essentially the same qpo frequencies .
we propose that this multi - valued qpo frequency and mass accretion rate correlation indicates the existence of many similar regimes of the accretion disk .
these regimes , with a very similar energy spectrum and qpo frequency , are distinguished from each other by the mass accretion rate or the total x - ray flux .
the ncos during the burst can be made almost perfectly coherent by taking into account a large @xmath1 .
this strongly suggests that this frequency is related to the neutron star spin frequency .
the large @xmath1 is attributable to the expansion or contraction of the neutron star photosphere during the burst . |
in other work @xcite two dimensional fields of quantum reference frames were described that were based on different quantum theory representations of the real numbers . because the description of the fields does not seem to be understood , it is worthwhile to approach a description in a way that will , hopefully , help to make the fields better understood .
this the goal of this contribution to the third feynman conference proceedings .
the approach taken here is based on three main points : * there are a large number of different quantum theory representations of natural numbers , integers , and rational numbers as states of qubit strings .
these arise from gauge transformations of the qubit string states . * for each representation , cauchy sequences of rational string states give a representation of the real ( and complex ) numbers .
a reference frame is associated to each representation . *
each frame contains a representation of all mathematical and physical theories .
each of these is a mathematical structure that is based on the real and complex number representation base of the frame .
this approach is summarized in the next section with more details given in the following sections .
as is well known , the large amount of work and interest in quantum computing and quantum information theory is based on quantum mechanical representations of numbers as states of strings of qubits and linear superpositions of these states .
the numbers represented by these states are usually the nonnegative integers or the natural numbers
. examples of integer representations are states of the form @xmath0 and their linear superpositions @xmath1 here @xmath2 is a state of a string of @xmath3 qubits where the qubit at location @xmath4 denotes the sign ( as @xmath5 ) and @xmath6\rightarrow \{0,1\}$ ] is a @xmath7 valued function on the integer interval @xmath8 $ ] where @xmath9 this last condition removes the redundancy of leading @xmath10 this description can be extended to give quantum mechanical representations of rational numbers as states of qubit strings @xcite and of real and complex numbers as cauchy sequences of rational number states of qubit strings @xcite . as will be seen , string rational numbers
can be represented by qubit string states @xmath0 where @xmath11 is a @xmath7 valued function on an integer interval @xmath12 $ ] with @xmath13 and @xmath14 and the sign qubit @xmath15 , at position @xmath4 .
a basic point to note is that there are a great many different representations of rational numbers ( and of integers and natural numbers ) as states of qubit strings .
besides the arbitrariness of the location of the qubit string on an integer lattice there is the arbitrariness of the choice of quantization axis for each qubit in the string .
this latter arbitrariness is equivalent to a gauge freedom for the choice of which system states correspond to the qubit states @xmath16 and @xmath17 .
this arbitrariness of gauge choice for each qubit is discussed in section [ gt ] in terms of global and local gauge transformations of rational string states of qubit strings .
a different representation of string rational numbers as states of qubit strings is associated with each gauge transformation . as will be seen in the next section
, there is a quantum representation of real numbers associated with each representation of the rational string states .
it is clear that there are a large number of different representations of real numbers as each representation is associated with a different gauge gauge representation of the rational string states . the gauge freedom in the choice of quantization axis for each qubit plays an important role in quantum cryptography and the transfer of quantum information between a sender and receiver .
the choice of axis is often referred to a reference frame chosen by sender and receiver for transformation and receipt of quantum information @xcite . here
this idea of reference frames is taken over in that a reference frame @xmath18 is associated with each quantum representation @xmath19 of real numbers .
since each real number representation is associated with a gauge transformation @xmath20 , one can also associate frames directly with gauge transformations , as in @xmath21 instead of @xmath22 it should be noted that complex numbers are also included in this description since they can be represented as an ordered pair of real numbers . or they can be built up directly from complex rational string states . from now on real numbers and their representations will be assumed to also include complex numbers and their representations .
an important point for this paper is that any physical or mathematical theory , for which the real numbers form a base or scalar set , has a representation in each frame as a mathematical structure based on the real number representation in the frame .
since this is the case for all physical theories considered to date , it follows that they all have representations in each frame as mathematical structures based on the real number representation in the frame .
it follows that theories such as quantum mechanics , quantum field theory .
qed , qcd , string theory , and special and general relativity all have representations in each frame as mathematical structures based on the real number representation associated with the frame .
it is also the case that , if the space time manifold is considered to be a @xmath23 of the real numbers , then each frame contains a representation of the space time manifold as a @xmath23 of the real number representation . to understand these observations better
it is useful to briefly describe theories . here
the usual mathematical logic @xcite characterization of theories as being defined by a set of axioms is used .
all theories have in common the logical axioms and logical rules of deduction ; they are distinguished by having different sets of nonlogical axioms .
this is the case whether the axioms are explicitly stated or not .
each theory described by a consistent set of axioms has many representations ( called models in mathematical logic ) as mathematical structures in which the theory axioms are true . depending on the axiom set the representations may or may not be isomorphic .
the real numbers axiomatized as a complete ordered field are an example of a simple theory . for this axiomatization
all representations of the real numbers are isomorphic .
however they are not the same .
all theories based on the real ( or complex ) numbers include the real ( and complex ) number axioms in their axiom sets .
these well known aspects of theories are are quite familiar in the application of group theory to physics .
each abstract group is defined by a set of axioms that consist of the general group axioms and additional ones to describe the particular group considered .
each group has many different representations as different mathematical systems .
these can be matrices or operators in quantum theory .
these are further classified by the dimensionality of the representation and whether they are reducible or irreducible .
the importance of different irreducible representations to describe physical systems and their states is well known . as sets of axioms and derived theorems , physical theories , unlike mathematical theories , also have representations as descriptions of physical systems and their properties .
it is immaterial here whether these representations are considered to be maps from theory statements to physical systems and their properties or from different mathematical representations of the theory to physical systems and their properties .
it is , however , relevant to note that theoretical predictions to be tested by experiment are or can be represented by equations whose solutions are real numbers .
spectra of excited states of nuclei , atoms , and molecules are examples .
the same holds for observables with discrete spectra such as spin , isospin , and angular momentum . here
the eigenvalues are the real number equivalents of integers or rational numbers .
it is useful to begin with a question `` what are the real numbers ? ''
the most general answer is that they are elements of any set of mathematical or physical systems that satisfies the real number axioms .
the axioms express the requirements that the set must be a complete , ordered field .
this means that the set must be closed under addition , multiplication and their inverses , a linear order must exist for the collection , and any convergent sequence of elements converges to an element of the set .
it follows that all sets of real numbers must at least have these properties .
however they can have other properties as well .
a study of most any mathematical analysis textbook will show that real numbers are defined as equivalence classes of either dedekind cuts or of cauchy sequences of rational numbers .. ] a sequence @xmath24 of rational numbers is a cauchy sequence if@xmath25 the proof that the set of equivalence classes of cauchy sequences are real numbers requires proving that it is a complete ordered field .
a similar situation holds for rational numbers ( and integers and natural numbers ) .
rational numbers are elements of any set that satisfies the axioms for an ordered field .
however the field is not complete .
the representation of rational numbers used here will be based on finite strings of digits or kits in some base @xmath26 along with a sign and @xmath27 point .
use of the string representation is based on the fact that physical representations of rational numbers ( and integers and natural numbers ) are given as digits or states of strings of kits or qukits ( with a sign and @xmath27 point for rational numbers ) in some base @xmath28 .
such numbers are also the base of all computations , mainly because of the efficiency in representing large numbers and in carrying out arithmetic operations .
the usefulness of this representation is based on the fact that for any base @xmath29 these string numbers are dense in the set of all rational numbers . here , to keep things simple ,
representations will be limited to binary ones with @xmath30 the representations will be further restricted here to be represented as states of finite strings of qubits .
this is based on the fact that quantum mechanics is the basic mechanics applicable to all physical systems .
the cauchy condition will be applied to sequences of these states to give quantum theory representations of real numbers .
it should be noted that there are also quantum theory representations of real numbers that are not based on cauchy sequences of states of qukit strings .
besides those described in @xcite there are representations as hermitian operators in boolean valued models of zf set theory @xcite and in a category theory context @xcite .
these representations will not be considered further here because of the limitation here of representations to those based on finite strings of qubits . here
a compact representation of rational string states is used that combines the location of the @xmath31 point and the sign .
for example , the state @xmath32 is a state of eight @xmath7 qubits and one @xmath33 qubit representing a rational string number @xmath34 in the ordinary decimal form .
qubit strings and their states can be described by locating qubits on an integer lattice rational string states correspond to states of qubits occupying an integer interval @xmath35.$ ] here @xmath36 , @xmath37 is the location of the @xmath33 qubit , and the @xmath7 qubits occupy all positions in the interval @xmath35.$ ] note that two qubits , a sign one and a @xmath7 one , occupy position @xmath38 for fermionic qubits this can be accounted for by including extra variables to distinguish the qubit types .
one way to describe rational string states is by strings of qubit annihilation creation ( ac ) operators @xmath39 acting on a qubit vacuum state @xmath40 also present is another qubit type represented by the ac operators @xmath41 .
here @xmath42 , and @xmath43 are integers . for this work
it is immaterial whether the ac operators satisfy commutation relations or anticommutation relations : @xmath44= \delta_{j , j^{{\prime}}}\delta_{{\alpha},{\alpha^{{\prime } } } } \\
\mbox{$[{a^{\dag}_{{\alpha},j}},{a^{\dag}_{{\alpha}^{{\prime}},j^{{\prime}}}}]=[{a_{{\alpha},j}},{a_{{\alpha}^{{\prime}},j^{{\prime}}}}]=0$}\end{array}\]]or @xmath45 with similar relations for the @xmath46 operators .
the @xmath46 operators commute with the @xmath47 operators .
rational number states are represented by strings of @xmath47 creation operators and one @xmath46 creation operator acting on the qubit vacuum state @xmath16 as @xmath48 here @xmath49 and @xmath50 with @xmath51 integers , and @xmath52\rightarrow \{0,1\}$ ] is a @xmath53 valued function on the integer interval @xmath12 $ ] .
alternatively @xmath11 can be considered as a restriction to @xmath12 $ ] of a function defined on all the integers .
an operator @xmath54 can be defined whose eigenvalues correspond to the values of the rational numbers one associates with the string states .
@xmath54 is the product of two commuting operators , a sign scale operator @xmath55 , and a value operator @xmath56 one has@xmath57 the operator is given for reference only as it is not used to define arithmetic properties of the rational string states .
there is a large amount of arithmetical redundancy in the states .
for instance the arithmetic properties of a rational string state are invariant under a translation along the integer axis .
this is a consequence of the fact that these properties of the state are determined by the distribution of @xmath58 relative to the position @xmath37 of the sign and not on the value of @xmath37 .
the other redundancy arises from the fact that states that differ in the number of leading or trailing @xmath59 are all arithmetically equal .
these redundancies can be used to define equivalence classes of states or select one member as a representative of each class . here the latter choice will be made in that rational number states will be restricted to those with @xmath60 for the sign location and those with @xmath11 restricted so that @xmath61 if @xmath62 and @xmath63 if @xmath64 this last condition removes leading and trailing @xmath10 the state @xmath65 is the number @xmath4 . for ease in notation from now on the variables
@xmath66 will be dropped from states . thus states @xmath67
will be represented as @xmath0 with the values of @xmath68 included in the definition of @xmath11 .
there are two basic arithmetic properties , equality , @xmath69 and ordering @xmath70 arithmetic equality is defined by @xmath71 here @xmath72 the set of integers @xmath73 for which @xmath74 and similarly for @xmath75 that is , two states are arithmetically equal if one has the same distribution of @xmath58 relative to the location of the sign as the other .
arithmetic ordering on positive rational string states is defined by @xmath76 where@xmath77 the extension to zero and negative states is given by@xmath78 the definitions of @xmath79 can be extended to linear superpositions of rational string states in a straightforward manner to give probabilities that two states are arithmetically equal or that one state is less than another state .
operators for the basic arithmetic operations of addition , subtraction , multiplication , and division to any accuracy @xmath80 are represented by @xmath81 the state @xmath82}}{a^{\dag}}_{1,-\ell}|0\rangle.\ ] ] where @xmath83 } } = { a^{\dag}}_{0,0}{a^{\dag}}_{0,-1}\cdots{a^{\dag}}_{0,-\ell+1},$ ] is an eigenstate of @xmath54 with eigenvalue @xmath84 as an example of the explicit action of the arithmetic operators , the unitary addition operator @xmath85 satisfies @xmath86 where @xmath87 is the resulting addend state .
it is often useful to write the addend state as @xmath88 more details on the arithmetic operations are given elsewhere @xcite .
note that these operations are quite different from the usual quantum theory superposition , product , etc . operations .
this is the reason for the presence of the subscript a. the arithmetic operators can be used to define rational number properties of rational string states and their superpositions .
they can also be used to define the cauchy condition for a sequence of rational string states .
let @xmath89 be any sequence of rational string states . here for each @xmath90 @xmath91 and @xmath92
is a @xmath7 valued function from a finite integer interval that includes @xmath93 the sequence @xmath94 satisfies the cauchy condition if @xmath95 in this definition @xmath96 is the state that is the arithmetic absolute value of the arithmetic difference between the states @xmath97 and @xmath98 the cauchy condition says that this state is arithmetically less than or equal to the state @xmath80 for all @xmath99 greater than some @xmath100 .
it must be emphasized that this cauchy condition statement is a direct translation of eq .
[ cauchynos ] to apply to rational string states .
it has nothing to do with the usual convergence of sequences of states in a hilbert or fock space .
it is easy to see that state sequences which converge arithmetically do not converge quantum mechanically .
it was also seen in @xcite that the cauchy condition can be extended to sequences of linear superpositions of rational states .
let @xmath101 here @xmath102}}$ ] is a sum over all integer intervals @xmath12 $ ] and over all @xmath7 valued functions from @xmath12.$ ] from this one can define the probability that the arithmetic absolute value of the arithmetic difference between @xmath103 and @xmath104 is arithmetically less than or equal to @xmath80 by@xmath105 the sequence @xmath106 satisfies the cauchy condition if @xmath107 where@xmath108 here @xmath109 is the probability that the sequence @xmath106 satisfies the cauchy condition .
cauchy sequences can be collected into equivalence classes by defining @xmath110 if the cauchy condition holds with @xmath111 replacing @xmath112 and @xmath113 replacing @xmath114 in eq .
[ cauchy ] . to this end
let @xmath115 $ ] denote the equivalence class containing the cauchy sequence @xmath116 similarly @xmath117 if @xmath118 where @xmath119 is given by eqs .
[ pjml ] and [ limpjkl ] with @xmath120 replacing @xmath104 in eq .
[ pjml ] .
the definitions of @xmath121 can be lifted to definitions of @xmath122 on the set @xmath115 $ ] of all equivalence classes .
it can be shown @xcite that @xmath115 $ ] with these operations and relations is a representation or model of the real number axioms . in this sense
it is entirely equivalent to @xmath123 which is the real number component of the complex number base for the hilbert space containing the rational string number states that were used to define @xmath124.$ ] another representation of real numbers can be obtained by replacing the sequences @xmath125 by operators .
this can be achieved by replacing each index @xmath90 by the rational string state that corresponds to the natural number @xmath90 .
these are defined by @xmath0 where @xmath126 and @xmath127 where @xmath128 is the lower interval bound , for the domain of @xmath11 as a @xmath7 function over the integer interval @xmath12.$ ] in this case each sequence @xmath129 corresponds to an operator @xmath130 defined on the domain of natural number states .
one has @xmath131 where @xmath90 is the @xmath54 ( defined in eq .
[ defn ] ) eigenvalue of the state @xmath132 @xmath130 is defined to be cauchy if the righthand sequence in eq .
[ defo ] is cauchy .
one can also give a cauchy condition for @xmath130 by replacing the natural numbers in the definition quantifiers by natural number states .
one can repeat the description of equivalence classes of cauchy sequences of states for the operators to obtain another representation of real numbers as equivalence classes of cauchy operators .
the two definitions are closely related as eq . [ defo ] shows , and should be equivalent as representations of real number axioms .
this should follow from the use of eq .
[ defo ] to replace the left hand expression for the right hand expression in all steps of the proofs that cauchy sequences of rational string states satisfy the real number axioms .
the representation of rational string states as states of strings of qubits as in eq .
[ rastrst ] implies a choice of quantization axes for each qubit .
usually one assumes the same axis for each qubit where the axis is fixed by some external field .
however this is not necessary , and in some cases , such as quantum cryptography @xcite , rotation of the axis plays an important role .
in general there is no reason why the axes can not be arbitrarily chosen for each qubit .
this freedom of arbitrary directions for the axes of each qubit corresponds to the set of possible local and global gauge transformations of the qubit states .
each gauge transformation corresponds to a particular choice in that it defines the axis direction of a qubit relative to that of its neighbors . here
a gauge transformation @xmath20 can be defined as an @xmath133 valued function on the integers , @xmath134 @xmath20 is global if @xmath135 is independent of @xmath73 , local if it depends on @xmath73 .
the effect of @xmath20 on a rational string state @xmath0 is given by @xmath136 where@xmath137 these results are based on the representation of @xmath135 as @xmath138 arithmetic relations and operators transform in the expected way . for the relations one defines @xmath139 and @xmath140 by@xmath141 these relations
express the fact that @xmath142 if and only if @xmath143 and @xmath144 if and only if @xmath145 for the operation @xmath146 one defines @xmath147 by @xmath148 then @xmath149 as expected .
this is consistent with the definition of @xmath146 in eq .
[ addn ] as a binary relation .
similar relations hold for @xmath150 it follows from these properties that the cauchy condition is preserved under gauge transformations .
if a sequence of states @xmath106 is cauchy then the sequence @xmath151 is u - cauchy which means that it is cauchy relative to the transformed arithmetic relations and operations .
for example , if a sequence @xmath152 satisfies the cauchy condition , then the transformed sequence @xmath153 satisfies the u - cauchy condition : @xmath154 these definitions and considerations extend to the cauchy operators .
if @xmath130 is cauchy the the above shows that @xmath155 is u - cauchy .
however @xmath156 is not a cauchy operator in the original frame and @xmath130 is not cauchy in the transformed frame . to see that @xmath156 is not cauchy in the original frame it is instructive to consider a simple example .
first one works with the cauchy property for sequences of states . let @xmath157\rightarrow \{0,1\}$ ] be a @xmath7 function from the set of all integers @xmath158 where @xmath159 define a sequence of states@xmath160for @xmath161 the sequence is cauchy as @xmath162 for all @xmath163 however for any gauge transformation @xmath20 the sequence @xmath164 is not cauchy as expansion of the @xmath165 in terms of the @xmath166 by eq .
[ adu ] gives @xmath167 as a sum of terms whose arithmetic divergence is independent of @xmath37 . to show that @xmath156 is cauchy in the transformed frame if and only if @xmath130 is cauchy in the original frame one can start with the expression for the cauchy condition in the transformed frame : @xmath168 from eq .
[ defou ] one gets @xmath169 from eqs . [ addn ] and [ addnpla ] applied to @xmath170 and eqs .
[ addau ] and [ addaua ] one obtains @xmath171 use of@xmath172for the absolute value operator gives @xmath173 finally from eq .
[ = au ] one obtains @xmath174 which is the desired result .
thus one sees that the cauchy property is preserved in unitary transformations from one reference frame to another . as was done with the cauchy sequences and operators , the u - cauchy sequences or their equivalents , u - cauchy operators ,
can be collected into a set @xmath175 of equivalence classes that represent the real numbers .
this involves lifting up the basic arithmetic relations @xmath176 and operations @xmath177 to real number relations @xmath178 and operations @xmath179 and showing that @xmath175 is a complete , ordered , field .
it is also the case that for almost all gauge @xmath20 the real numbers in @xmath180 are orthogonal to those in @xmath181 in the following sense .
one can see that each equivalence class in @xmath181 contains a state sequence @xmath182}\rangle$ ] where @xmath11 is a @xmath7 valued function on the interval of all integers @xmath183 let @xmath20 be a gauge transformation with associated state sequence @xmath184}\rangle.$ ] both sequences satisfy their respective cauchy conditions .
however the overlap @xmath185}|u_{0}{\gamma}_{n},us_{[u ,-
n]}\rangle \rightarrow 0 $ ] as @xmath186 this expresses the sense in which @xmath181 and @xmath180 are orthogonal .
as has been seen , one can define many quantum theory representations of real numbers as cauchy sequences of states of qubit strings or as cauchy operators on the qubit string states .
the large number of representations stems from the gauge ( global and local ) freedom in the choice of a quantization axis for the qubit strings .
complex number representations are included either as ordered pairs of the real number representations or as extensions of the description of cauchy sequences or cauchy operators to complex rational string states @xcite .
it was also seen that for each gauge transformation @xmath20 the real and complex number representations @xmath187 are the base of a frame @xmath188 .
the frame @xmath188 also contains representations of physical theories as mathematical structures based on @xmath189 the work in the last two sections shows that the description of rational string states as states of finite strings of qubits given is a description in a fock space .
( a fock space is used because of the need to describe states of variable numbers of qubits and their linear superpositions in one space . )
the arithmetic operations @xmath190 on states of these strings are represented by fock space operators .
the properties of these operators acting on the qubit string states are used to show that the states represent binary rational numbers
. finally equivalence classes of sequences of these states or of operators that satisfy the cauchy condition are proved to be real numbers @xcite .
the essential point here is that the fock space , @xmath191 , and any additional mathematics used to obtain these results are based on a set @xmath192 of real and complex numbers . for example , superposition coefficients of basis states are in @xmath193 , the inner product is a map from pairs of states to @xmath193 , operator spectra are elements of @xmath193 , etc .
the space time manifold used to describe the dynamics of any physical representations of the qubit strings is given by @xmath194 .
it follows that one can assign a reference frame @xmath195 to @xmath123 and @xmath196 here @xmath195 contains all physical and mathematical theories that are represented as mathematical structures based on @xmath123 and @xmath193 .
however , unlike the case for the frames @xmath197 the only properties that @xmath123 and @xmath193 have are those based on the relevant axioms ( complete ordered field for @xmath123 ) .
other than that , nothing is known about how they are represented .
this can be expressed by saying that @xmath123 and @xmath193 are external , absolute , and given .
this seems to be the usual position taken by physics in using theories based on @xmath123 and @xmath193 .
physical theories are silent on what properties @xmath123 and @xmath193 may have other than those based on the relevant axioms . however , as has been seen , one can use these theories to describe many representations @xmath19 and @xmath198 and associated frames @xmath188 based on @xmath133 gauge transformations of the qubit strings .
as noted , for each @xmath199 @xmath188 contains representations of all physical theories as mathematical structures over @xmath200 . for these frames
one can see that @xmath19 and @xmath198 have additional properties besides those given by the relevant axioms .
they are also equivalence classes of cauchy sequences @xmath201 or cauchy operators @xmath202 fig .
[ rcst1 ] is a schematic illustration of the relation between frame @xmath195 and the frames @xmath188 .
only three of the infinitely many @xmath188 frames are shown .
the arrows indicate the derivation direction in that @xmath192 based theory in @xmath195 is used to describe , for each @xmath20 @xmath19 and @xmath198 that are the base of all theories in @xmath188 .
note that the frame @xmath203 with the identity gauge transformation is also included as one of the @xmath204 it is _ not _ the same as @xmath195 because @xmath205 is not the same as @xmath123 .
the above relations between the frames @xmath195 and @xmath188 shows that one can repeat the description of real and complex numbers as cauchy sequences of ( or cauchy operators on ) rational string states in each frame @xmath188 . in this case
the fock space representation , @xmath206 , used to describe the qubit string states in @xmath188 , is different from from @xmath191 in that it is based on @xmath200 instead of on @xmath207 however the two space representations are related by an isomorphism that is based on the isomorphism between @xmath123 and @xmath208 it is useful to examine this more .
first consider the states of a qubit at site j. for an observer in frame @xmath209 these states have the general form @xmath210 where @xmath211 and @xmath212 are complex numbers in @xmath193 .
let @xmath20 be an @xmath133 gauge transformation where @xmath213 is defined by @xmath214 where @xmath215
then the states @xmath216 and @xmath217 in frame @xmath195 are seen by an observer in @xmath188 as the states @xmath16 and @xmath17 respectively as the quantization axis is different .
a similar situation holds for states of qubit strings . to an observer in @xmath195
the state @xmath218 is different from @xmath219 however an observer in frame @xmath197 with a different set of quantization axes for each qubit , would represent the state @xmath218 as @xmath0 as to him it is the same state relative to his axis set as it is to the observer in @xmath195 for his axis set . the situation is slightly different for linear superpositions of basis states . to an observer in @xmath195 ,
the coefficients @xmath220 in the state @xmath221 represent abstract elements of @xmath193 .
the same observer sees that this state in @xmath188 is represented by @xmath222 where @xmath223 as elements of @xmath198 , represent the same abstract complex numbers as do @xmath224 however an observer in @xmath188 sees this same state as @xmath225 to him the real and complex number base of his frame is abstract and is represented by @xmath207 in general an observer in any frame sees the real and complex number base of his own frame as abstract and given with no particular representation evident .
however the observer in frame @xmath195 also sees that what to him is the abstract number @xmath226 is the number @xmath227 as and element of @xmath19 in frame @xmath204 these considerations also extend to group representations as matrices of complex numbers . if the element @xmath228 as an abstract element of @xmath229 is represented in frame @xmath195 by the matrix @xmath230 where @xmath231 are elements of @xmath193 , then , to an observer in @xmath209 @xmath232 is represented in frame @xmath188 by @xmath233 here @xmath234 , as elements of @xmath198 , correspond to the same abstract complex numbers as do @xmath235 however an observer in @xmath188 sees this representation as @xmath230 which is the same as the @xmath195 observer sees in his own frame .
following this line of argument one can now describe another generation of frames with each frame @xmath188 in the role of a parent to progeny frames just as @xmath195 is a parent to the frames @xmath188 as in fig .
[ rcst1 ] .
this is shown schematically in fig .
[ rcst2 ] .
again , only three of the infinitely many stage 2 frames emanating from each stage 1 frame are shown . here
something new appears in that there are many different paths to a stage 2 frame . for each path , such as
@xmath236 @xmath237 is the product @xmath238 of two gauge transformations where @xmath239 an observer in this frame @xmath240 sees the real and complex number frame base as abstract , and given .
to him they can be represented as @xmath207 an observer in @xmath241 sees the real and complex number base of @xmath240 as @xmath242 however an observer in @xmath195 sees the real and complex number base of @xmath240 as @xmath243 the subscript @xmath244 denotes the fact that relative to @xmath195 the number base of @xmath240 is constructed in two stages .
first the fock space @xmath191 is used to construct representations @xmath245 of @xmath123 and @xmath193 as @xmath246 cauchy sequences of states @xmath247 then in frame @xmath241 the fock space @xmath248 based on @xmath249 is used to construct the number representation base of @xmath240 as @xmath250 cauchy sequences @xmath251 of qubit string states in @xmath252 one sees , then , that , for each path leading to a specific stage 2 frame , there is a different two stage construction of the number base of the frame .
this view from the parent frame @xmath195 is the same view that we have as observers outside the whole frame structure .
that is , our external view coincides with that for an observer inside the parent frame @xmath195 .
the above description of frames emanating from frames for 2 stages suggests that the process can be iterated .
there are several possibilities besides a finite number of iterations exemplified by fig.[rcst2 ] for 2 iterations .
[ rcst3 ] shows the field structure for a one way infinite number of iterations .
here one sees that each frame has an infinite number of descendent frame generations and , except for frame @xmath209 a finite number number of ancestor generations .
the structure of the frame field seen by an observer in @xmath195 is the same as that viewed from the outside . for both observers
the base real and complex numbers for @xmath195 are seen as abstract and given with no structure other than that given by the axioms for the real and complex numbers .
there are two other possible stage or generation structures for the frame fields , two way infinite and finite cyclic structures .
these are shown schematically in figs .
[ rcst4 ] and [ rcst5 ] .
the direction of iteration in the cyclic field is shown by the arrows on the circle rather than example arrows connecting frames as in the other figures .
for both these frame fields each frame has infinitely many parent frames and infinitely many daughter frames .
there is no single ancestor frame and no final descendent frames .
the cyclic field is unique in that each frame is both its own descendent and its own ancestor .
the distance between these connections is equal to the number of iterations in the cyclic field .
these two frame fields differ from the others in that the structure seen from the outside is different from that for an observer in any frame .
there is no frame @xmath195 from which the view is the same as from the outside . viewed from the outside
there are no abstract , given real and complex number sets for the field as a whole .
all of these are internal in that an observer in frame @xmath253 at generation @xmath73 sees the base @xmath254 as abstract with no properties other than those given by the relevant axioms .
the same holds for the representations of the space time manifold .
viewed from the outside there is no fixed abstract space time representation as a 4-tuple of real numbers associated with the field as a whole .
all space time representations of this form are internal and associated with each frame .
this is based on the observation that the points of a continuum space time as a 4-tuple of representations of the real numbers are different in each frame because the real number representations are different in each frame .
also , contrary to the situation for the fields in figs .
[ rcst1]-[rcst3 ] , there is no representation of the space time points that can be considered to be as fixed and external to the frame field .
the lack of a fixed abstract , external space time manifold representation for the two - way infinite and cyclic frame fields is in some ways similar to the lack of a background space time in loop quantum gravity @xcite .
there are differences in that in loop quantum gravity space is discrete on the planck scale and not continuous @xcite .
it should be noted though that the representation of space time as continuous is not a necessary consequence of the frame fields and their properties .
the important part is the real and complex number representation base of each frame , not whether space time is represented as discrete or continuous .
it is useful to summarize the views of observers inside frames and outside of frames for the different field types .
for all the fields except the cyclic ones an observer in any frame @xmath253 at stage @xmath73 sees the real number base @xmath255 of his frame as abstract and external with no properties other than those given by the axioms for a complete ordered field .
the observer also can not see any ancestor frames .
he / she can see the whole frame field structure for all descendent frames at stages @xmath256 except as noted below , the view of an outside observer is different in that he / she can see the whole frame field structure .
this includes the point that , to internal observers in a frame , the real and complex number base of the frame is abstract and external . for frame fields with a fixed ancestor frame @xmath195 , figs .
[ rcst1 ] , [ rcst2 ] , [ rcst3 ] , the view of an outside observer is almost the same as that of an observer in frame @xmath195 .
both see the real and complex number base of @xmath195 as abstract and external .
both can also see the field structure for all frames in the fields .
however the outside observer can see that frame @xmath195 has no ancestors .
this is not available to an observer in @xmath195 as he / she can not see the whole frame field .
the cyclic frame field is different in that for an observer in any frame at stage @xmath257 frames at other stages are both descendent and ancestor frames .
this suggests that the requirement that a frame observer can not see the field structure for ancestor frames , but can see it for descendent frames , may have to be changed , at least for this type of frame field .
how and what one learns from such changes are a subject for future work .
so far , frame fields based on different quantum mechanical representations of real and complex numbers have been described .
each frame contains a representation of physical theories as mathematical structures based on the real number representation base of the frame .
the representations of the physical theories in the different frames are different because the real ( and complex ) number representations are different .
they are also isomorphic because the real ( and complex ) number representations are isomorphic .
the description of the frame field given so far is incomplete because nothing has been said about the relation of the frame field to physics .
so far the only use of physics has been to limit the description of rational number representations to quantum mechanical states of strings of qubits .
the main problem here is that to date all physical theories make no use of this frame field .
this is evidenced by the observation that the only properties of real numbers relevant to physical theories are those derivable from the real number axioms for a complete , ordered field .
so far physical theories are completely insensitive to details of different representations of the real numbers .
this problem is also shown by the observation that there is no evidence of this frame structure and the multiplicity of real number representations in our view of the existence of just one physical universe with its space time manifold , and with real and complex numbers that can be treated as absolute and given .
there is no hint , so far , either in physical theories or in properties of physical systems and the physical universe , of the different representations and the frame fields .
this shows that the main problem is to reconcile the great number of different representations of the real and complex numbers and the @xmath194 space time manifold as bases for different representations of physical theories with the lack of dependence of physical theories on these representations and no evidence for them in our view of the physical universe .
one possible way to do this might be to collapse the frame field to a smaller field , ideally with just one frame . as a step in this direction one could limit quantum theory representations of rational string numbers to those that are gauge invariant .
this would have the effect of collapsing all frames @xmath253 at any stage @xmath73 into one stage @xmath73 frame is a gauge transformation .
it is not the @xmath258 element of one . ] .
the resulting frame field would then be one dimensional with one frame at each stage .
the idea of constructing representations that are gauge invariant for some gauge transformations has already been used in another context .
this is the use of the decoherent free subspace ( dfs ) approach to quantum error correction .
this approach @xcite is based to quantum error avoidance in quantum computations .
this method identifies quantum errors with gauge transformations @xmath20 . in this case
the goal is to find subspaces of qubit string states that are invariant under at least some gauge @xmath20 and are preserved by the hamiltonian dynamics for the system .
one way to achieve this is based on irreducible representations of direct products of @xmath133 as the irreducible subspaces are invariant under the action of some @xmath20 . as an example
, one can show that @xcite the subspaces defined by the irreducible 4 dimensional representation of @xmath259 are invariant under the action of any global @xmath20 .
the subspaces are the three dimensional subspace with isospin @xmath260 , spanned by the states @xmath261 and the @xmath262 subspace containing @xmath263 the action of any global @xmath20 on states in the @xmath260 subspace can change one of the @xmath264 states into linear superpositions of all states in the subspace . but it does not connect the states in the @xmath260 subspace with that in the @xmath262 subspace .
it follows that one can replace a string of @xmath265 qubits with a string of @xmath90 new qubits where the @xmath16 and @xmath17 states of the @xmath258 new qubit correspond to any state in the respective @xmath260 and @xmath262 subspaces of the 4 dimensional representation of @xmath266 any state of the @xmath90 new qubits is invariant under all global gauge transformations and all local gauge transformations where @xmath267 this replacement of states of strings of @xmath265 qubits by states of strings of @xmath90 new qubits gives the result that , for all @xmath20 satisfying eq . [ ueq ] , the @xmath188 frames at any stage @xmath73 all become just one frame at stage @xmath73 .
however this still leaves many gauge @xmath20 for which the new qubit string state representation is not gauge invariant .
another method of arriving at a gauge invariant description of rational string states is based on the description of the kinematics of a quantum system by states in a hilbert space @xmath268 based on the @xmath133 group manifold .
details of this , generalized to all compact lie groups , are given in @xcite and @xcite .
in essence this extends the well known situation for angular momentum representations of states based on the manifold of the group @xmath269 to all compact lie groups . for the angular momentum case
the the action of any rotation on the states @xmath270 gives linear combinations of states with different @xmath37 values but with the same @xmath128 value .
the hilbert space spanned by all angular momentum eigenstates can be expanded as a direct sum @xmath271 where @xmath272 labels the irreducible representations of @xmath273 qubits can be associated with this representation by choosing two different @xmath128 values , say @xmath274 and @xmath275 then any states in the subspaces @xmath276 and @xmath277 correspond to the @xmath16 and @xmath17 qubit states respectively .
these states are invariant under all rotations .
extension of this construction to all finite qubit strings gives a representation of natural numbers , integers and rational numbers that is invariant under all @xmath269 gauge transformations .
this development can also be carried out for any compact lie group where the quantum kinematics of a system is based on the group manifold @xcite . in the case of @xmath133 eq .
[ hl ] holds with @xmath278 .
the momentum subspaces @xmath279 are invariant under all @xmath133 transformations . as in the angular momentum case one
can use this to describe states of qubits as @xmath280 that are @xmath133 invariant .
this construction can be extended to states of finite strings of qubits .
details of the mathematics needed for this , applied to graphs on a compact space manifold , are given in @xcite . in this way one can describe representations of rational string numbers that are @xmath133 gauge invariant for all gauge @xmath20 .
there is a possible connection between these representations of numbers and the ashtekar approach to loop quantum gravity .
the ashtekar approach @xcite describes @xmath281 valued connections on graphs defined on a @xmath282 space manifold where @xmath281 is a compact lie group .
the hilbert space of states on all graphs can be represented as a direct sum of spaces for each graph .
the space for each graph can be further decomposed into a sum of gauge invariant subspaces .
this is similar to the spin network decomposition first described by penrose @xcite .
the connection to qubit strings is made by noting that strings correspond to simple one dimensional graphs .
states of qubit strings are defined as above by choosing two @xmath128 values for the space of invariant subspaces as in eq .
[ invqub ] .
it is hoped to describe more details of this connection in future work .
implementation of this approach to reduction of the frame field one still leaves a one dimensional line of iterated frames .
the line is finite , fig .
[ rcst2 ] , one way infinite , fig .
[ rcst3 ] , two way infinite , fig .
[ rcst4 ] , or closed , fig .
[ rcst5 ] .
because the two way infinite and cyclic fields have no abstract external sets of real and complex numbers and no abstract external space time , it seems appropriate to limit consideration to them . here
the cyclic field may be the most interesting because the number of the iterations in the cycle is undetermined .
if it were possible to reduce the number to @xmath4 , then one would end up with a picture like fig .
[ rcst1 ] except that the @xmath123 and @xmath193 base of the frame would be identified in some sense with the gauge invariant representations described in the frame . whether this is possible , or even desirable , or not , is a problem left to future work .
another approach to connecting the frame field to physics is based on noting that the independence of physical theories from the properties of different real and complex number representations can be expressed using notions of symmetry and invariance .
this is that @xmath283 note that the gauge transformations apply not only to the qubit string states but also to the arithmetic relations and operations on the string states , eqs .
[ = au ] and [ addau ] , to sequences of states , to the cauchy condition eq .
[ cauchyu ] , and to the definition of the basic field operations on the real numbers .
the symmetry of physical theories under these gauge transformations suggests that it may be useful to drop the invariance and consider at the outset candidate theories that break this symmetry .
these would be theories in which some basic aspect of physics is representation dependent .
one approach might be to look for some type of action whose minimization , coupled with the above requirement of gauge invariance , leads to some restriction on candidate theories .
this gauge theory approach is yet another aspect to investigate in the future .
there are some points of the material presented here that should be noted .
the gauge transformations described here apply to finite strings of qubits and their states .
these are the basic objects . since these can be used to represent natural numbers , integers , and rational numbers in quantum mechanics , one can , for each type of number , describe different representations related by @xmath133 gauge transformations on the qubit string states . here
this description was extended to sequences of qubit string states that satisfied the cauchy condition to give different representations of the real numbers .
a reference frame was associated to each real and complex number representation .
each frame contains a representation of all physical theories as mathematical structures based on the real and complex number representation base of the frame .
if the space time manifold is considered to be a @xmath23 of the real numbers , then each frame includes a representation of space time as a @xmath23 of the real number representation .
if one takes this view regarding space time , it follows that for all frames with an ancestor frame , an observer outside the frame field or an observer in an ancestor frame sees that the points of space time in each descendent frame have structure as each point is an equivalence class of cauchy sequences of ( or a cauchy operator on ) states of qubit strings .
it is somewhat disconcerting to regard space time points as having structure .
however this structure is seen only by the observers noted above .
an observer in a frame @xmath195 does not see his or her own space time points as having structure because the real numbers that are the base of his frame do not have structure .
relative to an observer in frame @xmath195 , the space time base of the frame is a manifold of featureless points .
it should also be noted that even if one takes the view that the space time manifold is some noncompact , smooth manifold that is independent of @xmath284 one still has the fact that functions from the manifold to the real numbers are frame dependent in that the range set of the functions is the real number representation base of the frame .
space time metrics are good examples of this . as is well known they play an essential role in physics . in quite general terms
, this work is motivated by the need to develop a coherent theory that treats mathematics and physics together as a coherent whole @xcite .
it may be hoped that the approach taken here that describes fields of frames based on different representations of real and complex numbers will shed light on such a theory .
the point that these representations are based on different representations of states of qubit strings shows the basic importance of these strings to this endeavor .
finally it should be noted that the structure of frames emanating from frames has nothing to do with the everett wheeler view of multiple universes @xcite .
if these multiple universes exist , then they would exist within each frame in the field .
this work was supported by the u.s .
department of energy , office of nuclear physics , under contract no .
w-31 - 109-eng-38 . | in this paper fields of quantum reference frames based on gauge transformations of rational string states are described in a way that , hopefully , makes them more understandable than their description in an earlier paper . the approach taken here
is based on three main points : ( 1 ) there are a large number of different quantum theory representations of natural numbers , integers , and rational numbers as states of qubit strings .
( 2 ) for each representation , cauchy sequences of rational string states give a representation of the real ( and complex ) numbers . a reference frame
is associated to each representation .
( 3 ) each frame contains a representation of all mathematical and physical theories that have the real and complex numbers as a scalar base for the theories .
these points and other aspects of the resulting fields are then discussed and justified in some detail . also
two different methods of relating the frame field to physics are discussed . |
measurement of the interstellar radiation field ( isrf ) , comprised of the direct - stellar and diffuse background radiation fields , is of considerable interest because the isrf , in particular at the far - ultraviolet ( fuv ) waveband , controls the physics and chemistry of the interstellar medium ( ism ) .
the fuv isrf is often expressed by a simple approximation referred to as the habing radiation field @xcite and was measured for the first time by @xcite with two geiger counters on board an aerobee rocket .
@xcite obtained another simple analytic expression that appeared to be in good agreement with all the previous results at that time , and it has been referred to as the draine s `` standard '' uv background . @xcite computed a theoretical model and tabulated the average uv isrf scaled to fit the radiation field observed by the s2/68 telescope aboard the td-1 satellite .
@xcite redetermined the isrf by fitting the spectrum derived by @xcite to power - law and scaling up by 15% to bring it into better agreement with the measurement of @xcite and extended the calculation to cover wavelengths from 0.09 @xmath1 m to 1000 @xmath1 m .
the majority of the diffuse fuv background radiation or the diffuse galactic light is generally believed to be of galactic origin , starlight scattered off by interstellar dust .
observations of the diffuse background then give us information on the scattering properties of the dust grains .
however , early attempts to measure the diffuse fuv background were motivated by the hope that a large fraction of the radiation in this waveband would be extragalactic ( see reviews by @xcite , @xcite , and @xcite )
. there had been serious disagreements among observers regarding the spatial and spectral variabilities of the fuv background radiation .
discrepancies among early measurements of the diffuse fuv radiation have lead to different interpretations as to its dominant contributor , either extragalactic light or galactic light scattered off dust grains .
a correlation between the fuv background and galactic neutral hydrogen column density ( @xmath0 ) was presented for the first time by @xcite , thereby implying that most of the fuv background is galactic in origin , consistent with scattering of the integrated fuv radiation field by interstellar dust .
substantial evidence supporting the conclusion has been reported later @xcite .
some studies , however , have suggested that the correlation with galactic latitude is more significant than that with galactic column density @xcite , or that the fuv background intensity shows only a weak correlation with galactic observables @xcite .
however , @xcite confirmed the correlation with galactic column density after subtraction of the inverse sin@xmath3 dependence from the data obtained with the d2b - aura satellite .
@xcite found that their data taken with the berkeley uvx spectrometer was three times better in dependence with galactic column density than with galactic latitude .
@xcite , @xcite , and @xcite showed also that a correlation exists between the diffuse infrared background intensity at 100 @xmath1 m as measured by the infrared astronomical satellite ( iras ) and the diffuse background intensity in the fuv wavelengths . @xcite demonstrated a similarity between the spatial power spectra of the iras 100 @xmath1 m cirrus images and the diffuse fuv background images obtained with the far ultraviolet space telescope ( faust ) . in the scale of a single cloud , @xcite discovered a galactic cirrus cloud , g251.2 + 73.3 , near the north galactic pole in fuv waveband and obtained a good correlation between the fuv intensity and iras 100 @xmath1 m surface brightness .
the existence of these properties thus provides strong observational support for the dust - scattered origin of most , if not all , of the fuv background .
different contributors to the diffuse fuv background have been suggested as well .
@xcite proposed fluorescent h@xmath4 emission as a significant contributor to the observed fuv background in addition to scattering by interstellar dust grains .
@xcite showed that the reported correlations could be explained by a combination of starlight scattered by interstellar dust and the h@xmath4 fluorescence .
later , the h@xmath4 fluorescence emission lines from the diffuse ism were observed with the berkeley uvx shuttle spectrometer @xcite and then by spear @xcite .
those measurements make it clear that the molecular hydrogen emissions are not the dominant contributor of the diffuse fuv background .
@xcite suggested that emission lines from a hot ( @xmath5 k ) galactic coronal gas would contribute to the diffuse fuv background , although this source would not be of sufficient intensity to influence the measurements of diffuse fuv continuum background .
@xcite estimated the possible contribution of a warm ( @xmath6 k ) ionized medium ( wim ) to the fuv background and attributed a fuv continuum intensity of @xmath7 photons @xmath8 s@xmath9 sr@xmath9 @xmath9 ( continuum unit , hereafter cu ) to the two - photon continuum emission from the wim ( see also @xcite ) .
@xcite showed that the two - photon emission from the wim accounts for approximately 20% of the diffuse fuv intensity near 1600 .
disagreements on the spectral shape of the diffuse fuv continuum background are also noticeable .
@xcite measured the diffuse fuv background at the north and south galactic pole regions , using a spectrometer flown on _ apollo _ 17 , and reported a spectrum that is flat between 1300 and 1525 and declining in intensity toward 1625 . the observations by @xcite and @xcite supported the decrease of intensity with increasing wavelength at high latitudes , using a spectrometer on board the soviet spacecraft `` prognoz-6 '' .
they also reported the flat spectrum in the intermediate latitudes and a rapid increase of the intensity with wavelength near and inside the milky way .
however , @xcite reported some evidence for a sharp rise in intensity longward of 1680 with instruments flown on an _ aries _ sounding rocket . the data observed by @xcite also showed a rise longward of about 1500 .
@xcite found a spectrally flat background over the wavelength band from 1200 to 1700 with the johns hopkins uvx experiment on board the space shuttle .
the spectrum obtained at a high galactic latitude with the berkeley uvx fuv spectrometer is relatively flat between 1400 and 1850 @xcite .
such discrepancies would be better understood with spectrophotometric observations over large area of the sky .
however , only a few studies have mapped a substantial fraction of sky in the fuv waveband .
@xcite presented the spatial distribution of the fuv stellar radiation at four passbands centered at 1565 , 1965 , 2365 , and 2740 from observations made with the s2/68 sky - survey telescope .
the first coarse map of the diffuse fuv radiation was presented by @xcite and later updated by @xcite using the data obtained with the elz spectrophotometer on board the d2b - aura satellite ( see also , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the most recent fuv band map is from the narrowband ultraviolet imaging experiment for wide - field surveys ( nuviews ; @xcite ) , which mapped one - quarter of the sky in fuv radiation ( @xmath10 , @xmath11 fwhm ) with @xmath12 imaging resolution , including features such as the upper scorpius region .
spectroscopy of plasma evolution from astrophysical radiation instruments ( spear , also known as far ultraviolet imaging spectrograph or fims ) provided the first large - area spectral sky survey of diffuse fuv radiation @xcite .
the mission observed @xmath13 of the sky and conducted deep pointed observations toward selected targets , such as supernova remnants @xcite , superbubbles @xcite , photoionized regions @xcite , and h@xmath4 molecular clouds @xcite .
the emission lines from multiply - ionized carbon , oxygen , silicon , and aluminum toward the north ecliptic and galactic pole regions observed with spear / fims were presented in detail by @xcite and @xcite . in this paper , we describe some general results on the diffuse fuv ( 13701710 ) continuum background observed with spear / fims . emission line sky - maps obtained with spear / fims is described in @xcite and edelstein et al .
( 2011 , in preparation ) .
this paper is organized as follows . in
2 we describe the observation and data reduction .
the fuv intensity maps , comparison of the diffuse fuv background with a model of the stellar radiation field are shown in 3 .
spectral shapes of the diffuse fuv background is described in 4 .
we compare the diffuse fuv continuum map with maps of @xmath0 , h@xmath2 , 100 @xmath1 m , and soft x - ray ( 1/4 kev ) in 5 . a discussion is presented in 6 .
finally , we conclude with a summary of results in 7 .
spear / fims is a dual - channel fuv imaging spectrograph ( `` short '' wavelength channel [ s - band ] 9001150 , `` long '' wavelength channel [ l - band ] 13501750 ; @xmath14 ) with a large field of view ( s - band , @xmath15 ; l - band , @xmath16 ) , designed to observe diffuse fuv emission lines .
spear is the primary payload on the first korean science and technology satellite , _
stsat-1 _ , and was launched into a @xmath17 km sun - synchronous orbit on 2003 september 27 .
the spear / fims mission , the instruments , their on - orbit performance , and the basic processing of the instrument data are described in detail by @xcite .
the spectrographs consist of the gratings , order baffling , the detector , a shutter - slit unit , a mirror unit with field baffling , and a deployable dust cover . the shutter can be selected to admit 0% , 1% , 10% , or 100% of the available light for safe and photometric observations of bright sources .
the instrument provides for the first time accurate determination of the fuv background spectral map covering a substantial fraction of sky .
the first galactic map of diffuse fuv ( l - band ) background measured by spear / fims was presented in @xcite . because s - band lacks the ly@xmath2-rejection filter that l - band has , the major continuum source for s - band is the instrumental background @xcite .
in addition , s - band has lower sensitivity than l - band , as well as a higher non - astrophysical background , including a strong ly@xmath18 ( 1026 ) airglow line .
therefore , in the present study , we analyze only the l - band data . given the wavelength dependence of the dust albedo and the fact that the composite stellar spectrum , which is dominated by ob - type stars , peaks near l - band , we expect that the astrophysical continuum in s - band is no higher than in l - band .
spear / fims observed the sky during the all - sky survey observation mode by scanning the entire sky along the short axis of the slit , i.e. , along the @xmath19 field direction of the @xmath20 fov . in each orbit , the fov scanned @xmath21 , from an ecliptic pole to the opposite ecliptic pole , during the eclipse time ( @xmath22 25 min ) , and the pole - to - pole scanning drifts @xmath23 along the ecliptic equator for one year because of the properties of a sun - synchronous orbit .
the fuv sky maps for l - band were created by binning photon and exposure events using the _ healpix _ tessellation scheme @xcite with @xmath24 pixels ( corresponding to the resolution parameter @xmath25 ) , unless otherwise specified , and l - band wavelength bins of 1.0 .
only the data from 1370 to 1710 , excluding the intense airglow line at 1356 , were used for the present analysis .
the data were selected for attitude - knowledge quality by using times when the derived attitude error is @xmath26 and contain @xmath27 and @xmath28 photons before and after removal of locally bright stars , respectively .
the diffuse background map was obtained by eliminating photon and exposure events in the locally intense pixels ( i.e. , stars ) from the starting sky map with @xmath29 pixels ( @xmath30 ) .
each orbital scan was made into an image and convolved with an edge detection kernel shaped like the stellar point spread function .
any region with counts 3@xmath31 above the local background was considered a star .
all photon events within a @xmath32 radius of the star centroid were flagged in the data .
this strategy effectively removed most faint and all bright stars ( edelstein et al .
2011 , in preparation ) . the remaining photon and exposure events
were binned into a map with @xmath24 pixels ( @xmath33 .
we repaired orbit number mismatches that were found after the publication of @xcite , and included about 100 more orbits other than those used in @xcite .
sky exposure was derived by generating time - marked exposure events and mapping the events to a sky position as detailed in @xcite .
the corresponding sky exposure map for the l - band with @xmath24 pixels ( @xmath25 ) , shown in figure [ exposure_map ] , covers @xmath34% of the sky and includes features such as deep exposures ( @xmath35 ks ) toward calibration and pointed study fields .
about 15% of these observations were taken using the 10% shutter position .
the exposures taken with the 10% shutter aperture were scaled by 0.1 , as determined in
[ sub : instrumental - background ] , and added to the map taken with the 100% shutter position . in the figure
, the gray color represents locations with zero exposure . in l - band , there are several spectral features , such as atomic emission lines ( @xmath361398 , @xmath361532 , @xmath371548 , 1551 , and @xmath361671 ) , quasi - band structures of h@xmath4 fluorescence lines , and two - photon continuum emission . since the physical processes that give rise to the emission lines are different than dominant continuum scattering ,
the variation of the emission lines may be independent from the dust - scattering processes that provide the bulk of the signal . if this emission is included when estimating the `` continuum '' intensity , the ability to compare the true fuv continuum background and other ism tracers is weakened , as is any discussion of the statistical properties of the fuv continuum . to obtain the true intensity of the fuv continuum , we may have to model all possible emission lines and continuum using spectral fits , as in @xcite .
however , the signal - to - noise ratio of the data with @xmath24 pixels is typically only @xmath38 ( as described in
[ sub : statistics ] ) , which is too small to provide reliable fits . to obtain reliable fits , large pixel sizes of @xmath39 (
@xmath40 were required , especially at high galactic latitudes , depending on locations in the sky ( see also edelstein et al .
2011 , in preparation ) . in this paper , instead of using spectral fits and variable pixel sizes , we excluded the wavelength regions of the strongest emission lines ( , , and , as shown in figures [ spec_vs_lon ] and [ spec_vs_lat ] ) and averaged the data in the wavelength regions of 13101520 , 15601660 , and 16801710 . unless otherwise specified , the fuv intensities in this paper were obtained in this way .
we report how the emission lines contribute to the total intensity of l - band .
edelstein et al . ( 2011 , in preparation ) showed that , on average , all atomic emission lines and h@xmath4 fluorescence lines contribute about 5.1% and 2.9% ( respectively ) to the intensity of spear / fims l - band .
our independent modeling using spectral fits indicated that the maximum contribution of emission lines is about 10% .
notably , the present results are not significantly different from results estimated using the whole l - band , which implies that the emission lines do not play a significant role in analyzing the general properties of the fuv continuum background .
we therefore use the terms `` diffuse fuv background '' and `` fuv continuum background '' interchangeably .
the contribution of two - photon emission to the diffuse fuv continuum background is discussed in
[ sub : two - photon ] .
the spear / fims data are affected by instrumental ( or detector ) background due to cosmic rays , radioactive decay within the detector , and thermal charged particles entering the instrument .
these background sources are relatively constant in time and uniform across the face of the detectors , implying that the instrumental background is spectrally flat intrinsically @xcite .
since the instrumental background is independent of the size of the shutter aperture while the astrophysical signal is not , the background can be determined by comparing the data observed with both 100% and 10% shutter apertures toward the same sky .
we thus determine the instrumental background rate assuming that the intensities measured with 100% and 10% shutter apertures are @xmath41 and @xmath42 , respectively . here ,
@xmath43 , @xmath44 , and @xmath45 denote the scale factor between two shutter apertures , the sky intensity measured with 100% shutter aperture , and the detector background , respectively .
we first estimate the detector background intensity averaged over the whole l - band instead of applying the method to every spectral bin , and then calculate the instrumental background spectrum in terms of cu .
we note that each pixel of a _ healpix _ map is hierarchically subdivided into many smaller pixels as the resolution parameter @xmath46 increases .
some of the subdivided pixels from a non - zero - exposure pixel at @xmath25 can have zero exposures , especially when the pixel is located at the boundary of each orbital scan .
hence , when comparing the 100% and 10% maps at @xmath25 , two coinciding pixels may not have the same sky coverage at higher resolutions .
this discrepancy can produce a large variation when comparing 100% and 10% maps with @xmath25 .
thus , we only used pixels ( @xmath25 ) with 100% sky coverage at the level of @xmath30 , i.e. , those pixels whose subdivided pixels at the level of @xmath30 all have non - zero exposures .
as shown in figure [ compare_100_10 ] , over the whole l - band , the averaged intensity of instrumental background that best fits the measurements is @xmath47 cu .
similar results that are consistent within @xmath38% were obtained using the higher or lower resolution parameters . in the fit
, we fully adopted errors in both axes and found the best - fit linear straight line between the two intensities @xcite . because the noise - property of the spear / fims data is dominated by photon - noise , we used only photon - noise in the present study .
the estimated instrumental background was also justified by an arbitrary increase or decrease of the background intensity and through visual inspection of the combined map . increasing or decreasing @xmath45 by more than @xmath22
50 cu caused the observed sky regions with the 10% aperture to appear unrealistically bright or faint in the combined map .
the scale factor between the two shutter apertures is @xmath48 as originally designed . denoting the instrumental background rate by @xmath49 counts s@xmath9 @xmath9 in the detector space , the instrumental background spectrum @xmath50 in units of cu in an @xmath51 spectral bin
is given by @xmath52 where @xmath53 , @xmath54 , @xmath55 , and @xmath56 denote the effective area at the wavelength @xmath57 , solid angle , wavelength binsize , and exposure time , respectively .
applying the maximum likelihood method to poisson data , the mean instrumental background intensity over the l - band is given by @xmath58 where @xmath59 is the number of wavelength bins .
we should note here that @xmath60 .
thus the instrumental background spectrum is obtained by @xmath61 and is shown in figure [ detector_spec ] . in the figure , @xmath62 error limits are also shown in dotted lines .
the spectrum of the instrumental background is inversely proportional to the effective area . using the solid angle of l - band @xmath54 of @xmath63 sr and average effective area @xmath64 of @xmath65 @xmath8
, we obtain the detector background rate @xmath66 counts s@xmath9 @xmath9 .
the instrumental background was independently estimated to be 0.020.04 counts s@xmath9@xmath9 by summing many shutter - closed dark exposures of 42 ks from 250 orbits @xcite .
@xcite adopted the same method and found a bit weaker detector dark background of about 0.01 counts s@xmath9@xmath9 .
we note that the instrumental background rate estimated in the present study is even smaller than the value obtained by @xcite .
the discrepancy is attributable to the light leakage in shutter - closed dark exposures .
the histogram of the sky exposure is shown in figure [ histograms](a ) on a logarithmic scale .
the dotted line in the figure shows the exposure distribution for the data observed with the 10% shutter aperture only .
figure [ histograms](b ) shows the intensity distributions of the observed total fuv intensity ( @xmath67 ) including both diffuse ( @xmath68 ) and direct stellar ( @xmath69 ) intensities , and the diffuse fuv background only .
the fuv background map has median and mean intensities of @xmath70 and of @xmath71 cu , respectively .
the reason for the difference between the median and mean values are obvious from the asymmetry of the intensity histogram .
the difference in the median value from @xcite is mainly due to subtraction of the instrumental background and inclusion of more orbits .
the total fuv intensity map has median and mean intensities of @xmath72 and @xmath73 cu , respectively .
the distribution of signal - to - noise ratio is also shown in figure [ histograms](c ) . for @xmath24 ( @xmath33 pixels ,
about 84% and 87% of pixels have significance of @xmath74 or higher from the diffuse fuv sky and the total sky , respectively . mean and median signal - to - noise ratios of @xmath68 are @xmath75 and @xmath38 , respectively , and the ratios of @xmath67 are @xmath76 and @xmath38 , respectively .
the statistical properties of the fuv intensities estimated over the whole l - band including the strong emission line regions are not significantly different .
the maps of the total intensity ( @xmath67 ) and the diffuse background intensity ( @xmath68 ) observed with spear / fims are shown in figure [ isrf_map ] and [ diffuse_map ] together with the signal - to - noise ratio maps .
as observed by @xcite and @xcite , the sky is strongly anisotropic in fuv .
not only is there a strong variation with galactic latitude , but also with longitude .
for example , much more fuv radiation is observed between longitude @xmath77 and @xmath23 than between @xmath78 and @xmath77 in the galactic plane .
gould s belt , inclined at @xmath79 to the galactic plane with the direction of tilt toward orion , shows up well in figure [ isrf_map ] ( @xcite ; @xcite and references therein ) .
obviously , the diffuse fuv background seems to follow direct starlight in general .
however , it will be shown in 3.3 and 5 that @xmath68 correlates better with dust , traced by 100 @xmath1 m emission , and column density rather than with the direct stellar intensity .
the largest diffuse intensity is toward the galactic plane and other regions where bright early - type stars coexist with suitable columns of interstellar dust , such as the obvious feature of the sco - cen association and the vela region .
we note that the sky is brightest at the vela region in both the total and diffuse background intensities .
we confirm the existence of a significant depression in both the diffuse and total fuv maps at high latitudes above less intense segments of the galactic plane from @xmath80 to @xmath81 , as noted in @xcite and @xcite .
@xcite showed that this direction corresponds to one of the most heavily reddened directions in the sky by comparing the fuv map with the reddening data of @xcite .
we also confirmed the strong reddening toward this direction with the more recent results given by @xcite , @xcite , and @xcite .
the dark region in @xmath82 and @xmath83 is in fact a molecular cloud referred to as aquila rift , which emits strong co line emission and has a maximum extinction of @xmath84 @xcite .
as can be seen in the orthographic projections of figure [ diffuse_map](b ) , the map reveals lots of detectable structures everywhere in the sky .
a full examination of the rich detail in the diffuse fuv map observed by spear / fims is beyond the scope of this paper .
however , it would be worthwhile to note that a large structure centered at @xmath85 is a feature produced by dust - scattered off starlight from a b1iii / iv + b2v spectroscopic binary @xmath2 vir with a distance of 87 pc @xcite .
an region surrounding the star has been detected by @xcite .
figures [ intensity_vs_lon ] and [ intensity_vs_lat ] examine the distribution of @xmath68 as a function of galactic coordinates .
the top panel of each figure plots a two - dimensional histogram in the space of intensity versus coordinates , while the bottom panel shows the median and the standard deviation of intensities from the median value within each of coordinate bins .
the sharp cutoff at low intensities in figure [ histograms](b ) and the shape of the latitude profile of the diffuse fuv emission in figure [ intensity_vs_lat ] resemble the corresponding features shown in the diffuse h@xmath2 emission ( figures 10 and 11 , respectively , in @xcite ) .
these similarities between the diffuse fuv and h@xmath2 emissions may be explained to some extent by a simple plane - parallel model . in the idealized model ,
the cosecant effect is removed by plotting @xmath86 versus @xmath87 .
this is examined in figure [ isinb_vs_sinb ] .
the top panel again plots a two - dimensional histogram and the bottom panel shows the median and standard deviation , with the trend for the northern galactic hemisphere split from that of the southern hemisphere
. it should be noted , however , that the plane - parallel model turns out to be only approximate in describing the diffuse fuv background , as is also the case for the gas and the wim @xcite .
the horizontal line in figure [ isinb_vs_sinb ] delineates the overall median value of @xmath88 cu .
the decrease in @xmath86 at @xmath89 ( i.e. , @xmath90 ) is attributed to a strong dust extinction in the galactic disk . the diffuse fuv intensity decreases from @xmath91 through 1.0 in both galactic hemispheres .
similar effects have been observed in the h@xmath2 and distributions , and the departure from the @xmath92 law was attributed at least in some extent to the presence of the local bubble @xcite , a localized region of very low - density gas around the sun , as pointed out by @xcite for . since the local bubble is believed to be elongated perpendicular to the galactic plane @xcite , a larger fraction of column density of the ism is carved out at high latitudes , yielding the apparent departure from the @xmath92 law . however , the main reason for the departure in the diffuse fuv background is due to the larger contribution of dust - scattered fuv radiation at low latitude , as shown for the h@xmath2 radiation by @xcite .
@xcite showed that simple addition of the predicted uv light from all of the stars listed in common star catalogs provides an estimate of the expected isrf in the uv band which is in good agreement with the more complex models ( see also , * ? ? ?
thus , we made a model of direct starlight using the _ hipparcos _ catalog @xcite for stellar location , distance , spectral type , and brightness .
catalog has a limiting magnitude of @xmath93 and was chosen for the calculation of three - dimensional locations of stars .
the model is basically the same as in @xcite , @xcite , and @xcite .
based on the spectral type of each star , we derived a temperature and effective gravity using the tables from @xcite and calculated a spectral energy distribution for each star with a new grid of kurucz models @xcite , rather than using the original grid of kurucz models .
the intrinsic fuv luminosity of each star was obtained by scaling the model flux with the factor derived from the formula , as in @xcite : @xmath94 where @xmath95 is the scaling factor corresponding to the stellar radius , @xmath96 the distance from the sun , @xmath97 ergs @xmath8 sec@xmath9 @xmath98 the absolute visual magnitude corresponding to zero magnitude @xcite , @xmath99 the model flux ( ergs @xmath8 sec@xmath9 @xmath98 ) at @xmath100 band filter , @xmath100 the observed johnson magnitude , and @xmath101 the optical depth at @xmath100 band filter which can be estimated from @xmath102 . here
, the absolute visual magnitude is calculated from the standard relation @xmath103 and @xmath104 , and the color excess @xmath105 obtained from the observed @xmath106 and the theoretical @xmath106 as derived from the kurucz model spectrum . in many cases
multiple or complex spectral types were associated with a single star ; in such cases , we used the first listed spectral type .
we then integrated the stellar flux attenuated by the dust distribution along given lines of sight .
the resulting stellar fluxes in every pixel are summed up to yield the `` stellar equivalent diffuse intensity ( sedi ) '' as defined by @xcite : @xmath107 where @xmath108 is the luminosity of each star at the 13701710 band ( excluding the strongest emission line regions ) in photons s@xmath9 @xmath9 , @xmath109 is the distance to the star in centimeters , @xmath110 is the optical depth to each star at the waveband , and @xmath111 is the solid angle ( steradians ) of a pixel .
the optical depth at the fuv wavelength is calculated by @xmath112 where @xmath113 and @xmath114 are the dust extinction cross - sections at the fuv wavelength and the v band filter for the average milky way dust with @xmath104 as from @xcite and @xcite .
we used the resolution parameter @xmath115 corresponding to an angular resolution of @xmath116 and the number of pixels @xmath117 in order to increase the number of pixels including stars .
we also calculated the sedi using the td-1 stellar catalog @xcite .
the td-1 satellite performed the first uv all - sky survey and measured the absolute uv fluxes of point sources down to the tenth visual magnitude for unreddened early b stars .
the td-1 catalog presents the stellar fluxes in four bands ( each 310330 wide , centered at 1565 , 1965 , 2365 , and 2740 ) for 31,215 stars .
the sedi of the td-1 stars is defined by @xmath118 using the stellar fluxes @xmath119 at the 1565 band .
@xcite also used the td-1 star catalog to model the diffuse fuv background , and they found that stars with 10 times higher flux than the td-1 catalog cutoff contribute about 21 cu to the fuv background .
therefore , the td-1 catalog is complete enough for the analysis of the diffuse fuv background .
first of all , the td-1 sedi is compared with the _ hipparcos _ sedi and the direct stellar intensity obtained from spear / fims data in figures [ compare_stellar](a ) and ( b ) , respectively .
the _ hipparcos _
sedi predicts the td-1 sedi very well in the statistical sense .
however , the _ hipparcos _ sedi is a bit higher than the td-1 sedi at intensities lower than @xmath120 cu , implying that the stars in the td-1 catalog are not complete down to the intensity level that was estimated with the _ hipparcos _ catalog , although the incompleteness is not so significant .
larger scatters in figure [ compare_stellar](b ) than those shown in figure [ compare_stellar](a ) are mainly due to the coarse algorithm used to identify stars in the spear / fims data analysis .
however , the correlation between the td-1 sedi and the spear / fims @xmath69 is good enough to move into the following arguments .
we note in figures [ compare_stellar](c ) and ( d ) that @xmath68 shows more or less a general correlation with the _ hipparcos _ and td-1 sedis .
however , as noted by @xcite , as long as one avoids the brightest early - type stars in the sky by more than a few degrees , there should not be a strong correlation between line - of - sight starlight and line - of - sight dust illumination .
in fact , we found that the correlation with the direct stellar intensity is rather weaker than that with the infrared 100 @xmath1 m emission and @xmath0 , as discussed in 5 .
it is also obvious that the @xmath67 is brighter than both sedis , as in figures [ compare_stellar](e ) and ( f ) , especially at relatively low intensities @xmath121 cu .
the deviation of @xmath67 from the _ hipparcos _ and td-1 sedis increases as the sedis decrease .
this trend is a direct result of the intensity distribution of @xmath68 as can be seen in figures [ compare_stellar](c ) and ( d ) , in which @xmath68 shows no strong correlation with the sedis .
more specifically , within each `` pixel '' ( @xmath116 ) , there is an excess of diffuse fuv , which is above the estimate from the stellar fluxes that are directly detected and/or modeled within that pixel and does not correlate with the direct stellar flux in the pixel . for low - intensity pixels , the excess of diffuse fuv would mainly be due to dust - scattered light from stars that are not located along the line of sight of the pixels .
the conclusion is discussed much more clearly in 4.2 , which attempts to link a harder spectrum at low intensities to this excess .
in fact , luminous uv stars produce extended dust - scattered halos that are far greater than the volume sample of the pixels here ( e.g. , * ? ? ? * ) .
we thus conclude that simply adding stellar flux underestimates isrf , especially at the low intensity of @xmath122 cu , against the results of @xcite and @xcite .
three - dimensional radiative transfer models may help to assert what portion of the isrf is the result of the dust - scattered fuv background .
it would be worthwhile to note that the contribution of unresolved faint stars to our diffuse fuv background map is negligible .
the diffuse background or the total intensity would be correlated at low intensity pixels with the model sedi if this was the case .
however , it is obvious in figure [ compare_stellar](c ) and ( d ) that this is not the case .
the difference of more than an order between @xmath123 and the mean value of @xmath67 at @xmath124 cu in figures [ compare_stellar](e ) and ( f ) is unlikely due to the contribution from unresolved stars .
we thus conclude that residual starlight is not a significant contributor to the diffuse fuv background .
the contribution from faint stars seen in figure [ compare_stellar](a ) is much smaller than the difference between @xmath123 and @xmath67 in figures [ compare_stellar](e ) and ( f ) .
many studies ( i.e. , * ? ? ?
* ; * ? ? ?
* ) have estimated the contribution of unresolved faint stars with several methods in their analysis of the fuv background data , and they all have concluded that the contribution is not significant
. additional support for this conclusion will be provided in 4 from an argument based on hardness ratio .
we also note that the large difference between the diffuse fuv intensity and the model sedis can not be due to h@xmath4 fluorescence lines and two - photon continuum emission , which may account for @xmath125% ( edelstein et al .
2011 , in preparation ) and @xmath126% ( see [ sub : two - photon ] ) of the diffuse intensity , respectively .
zodiacal light , sunlight scattered by interplanetary dust , is also negligible at these wavelengths since the solar g0 spectrum falls steeply at the fuv wavelengths @xcite , and we have confirmed that our data show no systematic trend with ecliptic latitude . scattering of diffuse ly@xmath2 radiation ( geocoronal / interplanetary ) as determined from laboratory measurements is also negligible .
the insignificance of unresolved faint stars , zodiacal light , and ly@xmath2 scattering is also justified from the agreement of the so - called `` isotropic component '' estimated in 5 with the previous measurements .
the direct stellar and diffuse background intensities should show strong anisotropy not only in intensity but also in spectral shape . in figures
[ spec_vs_lon ] and [ spec_vs_lat ] , the average spectra of the diffuse background , direct starlight , and total radiation in various galactic longitude and latitude ranges , respectively , are shown .
the direct stellar spectra were obtained by subtracting the diffuse background spectra from the total spectra in each coordinate range .
we note that the fuv intensities are strongest in @xmath127 and in @xmath128 , where the vela region is located , as was seen in figure [ diffuse_map ] .
the vela region is the brightest not only in emission lines due to the vela supernova remnant but also in the fuv continuum ( cf .
the average background intensity is about the same as the direct stellar intensity in all coordinate ranges shown in the figures , and the total intensity is approximately two times higher than the direct stellar intensity . ignoring the latitude range @xmath129 , where exposures and signal - to - noise ratios are small , the intensity ratio of the diffuse background to the direct starlight decreases with galactic latitude , and then increases at the highest latitude range of @xmath130 .
the trend is easily understandable , at least qualitatively , by adopting the results from a simplified spherical dust model with an embedded source .
the scattered to absorbed stellar flux ratio in the model increases monotonically with optical depth and can easily be of the order of unity or higher for optical depth of @xmath131 @xcite . the optical depth of @xmath132 at 1550 is estimated at @xmath133 @xmath8 using the standard gas - to - dust ratio @xcite and the milky way dust extinction cross - section @xcite .
therefore , the approximate equality between the diffuse and direct stellar intensities at the galactic plane is obtained .
the decrease of the diffuse to direct stellar intensity ratio with galactic latitude is also explained since the optical depth decreases with latitude . however , at the highest galactic latitude range @xmath130 , the `` isotropic component '' of @xmath134 cu discussed in 6 contributes significantly to the diffuse background in addition to the scattered radiation of galactic starlight in the latitude range , and thus causes the ratio to increase . a weak rise in the diffuse spectra longward of about 1550
is shown in most of the latitudes in figure [ spec_vs_lat ] , although the fuv continuum spectra are flat in general . however , the dependence of the rise on the galactic latitude is not clear .
the rise is also noticeable at @xmath135 in figure [ spec_vs_lon ] , where most of observations were performed at high latitudes as seen in figure [ exposure_map ] .
it should be noted as well that the softer the diffuse spectrum is , the softer the direct stellar spectrum also is . a much sharper rise longward of @xmath136 than that found here was obtained by @xcite , in which they concluded that the rise was due to very large numbers of faint stars contributing to the isrf .
later , @xcite reanalyzed the data and attributed the rise to an actual rise in the interstellar grain albedo
. however , these features may not be due to a single cause .
firstly , a stronger extinction at a shorter wavelength leads to the reddening or softening not only of direct stellar , but also of diffuse spectra , unless the dust albedo decreases with wavelength as will be shown in 6 ( figure [ aeffect ] ) . secondly ,
if there is no significant contribution from bright early - type stars near a given sight line , faint late - type stars should contribute , at least in part , to the spectra in addition to the dust - scattered light from distant early - type stars .
it should also , more importantly , be noted that the diffuse background spectrum is always softer than the direct stellar spectrum .
the rise of the dust albedo with wavelength is in fact closely related to the relative softness of the scattered to the direct stellar spectra , as will be discussed in 6 .
we also note that both the direct stellar and diffuse spectra are harder , or bluer , in the galactic longitude range @xmath127 , where the vela region dominates the radiation field , than other longitude ranges , implying the existence of lots of hot stars and/or less extinction in the region .
a relatively sharp rise longward of about 1690 at @xmath137 is also found mostly in the diffuse background spectrum . however , the origin of this rise is not clear at this time .
we now compare the spatially - averaged total ( direct stellar + diffuse ) fuv spectrum observed with spear / fims with the previous models and observations in figure [ isrf_compare ] .
it can be easily noted that all previous results are higher , up to a factor of more than 3 for the draine s model , than the spear / fims data . in figure
[ spec_vs_lat ] , the observed total spectrum is stronger only at @xmath128 than the draine s `` standard '' radiation field , while the intensities in other latitude ranges are in general weaker . the relative weakness of the observed data compared to the model isrfs indicates that the previous works are mostly biased toward bright regions in galactic plane . in figure [ isrf_compare ] , we also plotted the average spectra of the diffuse background and the direct starlight .
the figure shows once again that the direct starlight and the diffuse background contributes approximately the same amounts to the total spectrum on average , implying underestimation of the isrf when only direct starlights are added
. the spectral shapes of the model isrfs calculated by @xcite and the habing field are more or less similar to the spear / fims data , while the draine s `` standard '' and mathis background radiation fields are a bit harder than that of spear / fims .
the observed spectrum of @xcite is softer than the spear / fims data .
we also note that the diffuse background spectrum averaged over all the sky is a bit softer than the stellar spectrum , and that discrete spectral features such as atomic emission lines are mainly shown in the diffuse spectrum only , as already shown in figures [ spec_vs_lon ] and [ spec_vs_lat ] .
the relative softness or redness of the diffuse background compared to the direct stellar spectrum is once again attributed to rise of the dust - scattering albedo at a longer wavelength , as will be discussed in 6 . in order to further investigate spectral variations of the diffuse background
, we plotted a map of the hardness ratio defined as the ratio of average intensities at @xmath138 and at @xmath139 ( excluding @xmath140 to avoid the emission line ) in figure [ hardness_map ] . in the figure
, we selected a color scheme to represent the region with relatively blue or hard spectra as blue colored .
the spectral hardness map shows highly patched structures and no clear trend along with galactic latitudes .
instead , we note that the sky can be largely divided into two blocks mainly along galactic longitudes .
in fact , we smoothed the hardness ratio map using a gaussian kernel with full width at half maximum of @xmath141 and saturated the color range to clarify the division of the sky based on the hardness ratio .
a bit harder radiation is observed between longitude @xmath77 and @xmath23 , where stronger radiation is found in figure [ diffuse_map ] , than between @xmath78 and @xmath77 .
interestingly , we found that the hardness ratio of the diffuse background follows in general the spatial distribution of ob - type stars in the galactic plane . in figure [ td1_stars ] , we plotted maps for numbers of ob - type stars and of a - type stars using the td-1 stellar catalog .
we note from the figure that the galactic plane can be divided into three regions : @xmath142 where ob - type stars are predominantly found , @xmath143 and @xmath144 where a - type stars are dominant , and @xmath145 where both ob- and a - types are rare . in the galactic longitudes where ob - type stars are dominant , the hardness ratio is generally higher than the other regions .
we also plotted the ratio versus galactic longitude and latitude in figures [ hardness_vs_lon ] and [ hardness_vs_lat ] , respectively . in figure
[ hardness_vs_lon ] , the median value of the hardness ratio at each longitude bin is generally lower in @xmath146 than in @xmath147 .
on the other hand , figure [ hardness_vs_lat ] shows no clear trend along galactic latitude . in order to confirm the close relation between the hardness ratio and the number of ob - type stars
, we calculated the average hardness ratio and the number of ob - type stars within each of the @xmath148 longitude bins , and plotted the average hardness ratio versus the number of ob - type stars in figure [ hardness_obtype ] .
the figure shows a clear correlation between the average hardness ratio and the number of ob - type stars .
we also note a dip at @xmath135 in figure [ hardness_vs_lon ] .
the dip does not seem to be an artifact caused by a relatively low exposure at low latitudes in this longitude range . in the light of the argument described in the previous paragraph
, the softness in the longitude region would be a direct consequence of no bright stars in the range .
figure [ hardness_vs_intensity ] shows the hardness ratio versus the diffuse fuv intensity .
rises of the hardness ratio at high and low intensities are noted in the figure . to understand the trend ,
we calculated the hardness ratio versus the direct stellar intensity using the previous stellar model ; the result is shown in figure [ hardness_model ] . in the calculation of the hardness ratio
, we incorporated the wavelength dependence of dust extinction as well .
we note that the model ratios predict the observed values nicely at high intensities ( @xmath149 cu ) , of which most of the radiation is from low galactic latitudes .
therefore , we attribute the rise of the hardness ratio at high intensities ( @xmath149 cu ) mostly to the spectral hardness of stars that are located at low galactic latitudes . however , the slight rise of the hardness ratio at low intensities ( @xmath150 cu ) can not be explained by stellar types at high galactic latitudes .
one possible explanation for this rise is the contribution of dust - scattered starlight from earlier stars rather distant than nearby stars that are located along the same line of sight for each pixel .
this is supported by figures [ compare_stellar](c ) and ( d ) , which show clear excesses of the diffuse background above the nearby direct stellar intensity . in fact , our preliminary results of the monte carlo simulation , which will be presented in detail elsewhere , show that b - type stars are significant sources of the diffuse fuv background even at the high galactic latitudes where most stars are a - types .
@xcite also noted that b - type stars contribute the largest amount to the isrf in his simplified model .
we thus conclude that both the rise of the hardness ratio and the excess of the diffuse background over the direct stellar intensity at low intensities are largely due to the contribution of b - type starlight scattered off dust grains .
however , at the lowest intensities of less than a few hundred cu , the `` isotropic component '' would contribute significantly .
a detailed study on the relative significance between the two components may need extensive radiative transfer models that are beyond the scope of this paper .
we now examine correlations of the observed fuv background with various galactic quantities , not only with the column density and 100 @xmath1 m emission but also with the h@xmath2 and soft x - ray ( 1/4 kev ) emissions .
although extensive correlation studies between the diffuse fuv background , the column density , and 100 @xmath1 m emission have been performed , no detailed correlation study with other waveband observations has been done .
the all - sky map of column density are from the leiden / argentine / bonn ( lab ) survey @xcite .
the 100 @xmath1 m and h@xmath2 emission maps are from @xcite and @xcite , respectively . the soft x - ray ( 1/4 kev ) map is from the _ rosat _ all - sky survey ( rass ) map @xcite .
the data interpolated onto a healpix projection were obtained from the legacy archive for microwave background data analysis ( lambda ) .
correlations between the diffuse fuv intensity and other quantities are shown in figure [ correlation ] in log - log scales .
two - dimensional histogram images and contours are shown in the figure instead of the usual pixel - by - pixel comparison to clarify their correlation .
correlation coefficients estimated in logarithmic scale are also shown in figures [ correlation ] .
it is also worthwhile to note that the correlation coefficient between the diffuse fuv background and the direct stellar radiation is @xmath151 , much less than those for @xmath0 and @xmath152 . in figure [ correlation_linear ]
, we examine the correlations in linear - linear scales , especially at high galactic latitudes ( @xmath153 ) .
the general similarity in the latitude dependence of the data sets ( the column density , 100 @xmath1 m , h@xmath2 , and fuv emissions ) may dominate their correlations . to reduce the latitude dependence , we therefore calculated the average intensities multiplied by @xmath87 within each of the @xmath154 latitude intervals .
the correlations between the quantities multiplied by @xmath87 are shown in figure [ correlation_sinb ] , in which only the data sets observed at high galactic latitudes ( @xmath155 ) were examined , in order to minimize the dust extinction effect at low galactic latitudes.@xmath156 the numbers outside and inside the parentheses in figure [ correlation_sinb ] are correlation coefficients calculated in linear and logarithmic scales , respectively . here , we note that figure [ correlation_sinb ] shows strong correlations between emission tracers , implying that the correlations between the fuv background and other emission tracers are not simply due to the latitude dependence of the galactic observables .
correlation plots of the diffuse fuv background with column density and 100 @xmath1 m intensity are shown in figures [ correlation](a ) and ( b ) , respectively .
there is large scatter in the correlation plots .
we note that the diffuse fuv continuum intensity saturates at about @xmath157 @xmath8 , which was found for the first time by @xcite . the same effect of saturation is also shown in comparison with the 100 @xmath1 m intensity .
this saturation at a high intensity can be attributed to strong dust absorption of the fuv radiation at low galactic latitudes .
flattening at a low intensity , which has been often attributed to extragalactic background , is also seen in figures [ correlation](a ) and ( b ) .
as is obvious in the figures , the diffuse fuv intensity for a given @xmath0 or 100 @xmath1 m brightness varies more than by a factor of 23 , up to an order of magnitude , with directions in the sky .
@xcite argued that most of the observed variation is a result of the spatially , severely anisotropic , isrf .
since the diffuse fuv background depends not only on the interstellar dust but also on the in - situ stellar radiation field , the usual pixel - by - pixel comparison might be affected by the strong anisotropic stellar radiation field .
we thus averaged the fuv data corresponding to a given column density or a given 100 @xmath1 m brightness and compared the average fuv intensity with @xmath0 and 100 @xmath1 m brightness . in this way , the effect of the spatially varying stellar radiation field may be averaged , and only the effect due to or dust column density can be investigated . the average fuv intensity versus @xmath0 and @xmath152 are shown in figure [ correlation_linear ] , together with the best regression lines . in the figure ,
the data used in the fit are denoted by filled circles , and the data not used in the fit by hollow circles . in the fit
, we used only the data for @xmath153 .
the best - fit regression lines are given by @xmath158 and @xmath159 for column density and 100 @xmath1 m emission , respectively .
@xcite reported variations in the slopes and brightness axis intercepts corresponding to the various scans . as noted by @xcite , there is no unique , canonical relationship between background intensity and a hydrogen column density that is obeyed everywhere in the sky .
principally , this result is not surprising , since the detailed viewing geometry and the illuminating uv stellar radiation field are highly anisotropic , and several of the physical parameters , such as dust properties , are expected to vary from region to region in the ism .
however , it is obvious , as shown in figure [ correlation](a ) and ( b ) , that there is a general correlation in the global scale between the diffuse fuv background and both the neutral hydrogen column density @xmath0 and ir 100 @xmath1 m emission .
the linear correlation has been measured by many different investigators , and the results published before 1991 were summarized in @xcite as @xmath160 cu , which is consistent with our result .
the flaring and large scatters shown in these linear - linear correlation plots are in general due to large variations of the diffuse fuv intensity as noted in the log - log correlation plots of figure [ correlation ] .
note that the larger scatters are found at the higher mean intensities , the property that will be discussed in more detail in 6 .
we also note a straight - line correlation for @xmath161 mjy / sr in figure [ correlation_linear](b ) .
@xcite found that there was also a linear correlation between the optical surface brightness of high - latitude cirrus clouds and their 100 @xmath1 m intensity at least up to @xmath22 8 mjy / sr .
this is the range in which these objects remain optically thin and single scattering by dust grains is dominant .
the systematic deviation from a straight - line correlation for @xmath162 mjy / sr in figure [ correlation_linear](b ) would be due to the fact that at the higher @xmath163 values we are entering lower galactic latitudes where the effects of the strongly forward - directed scattering phase function of dust produces strongly enhanced fuv intensities compared to those produced by the scattering at mostly larger scattering angles at higher latitudes , where the 100 @xmath1 m intensities are generally low . in linear scale , the best regression line between @xmath68 and @xmath164 is given by @xmath165 where r denotes rayleigh ( 1 r @xmath166 photons @xmath8 s@xmath9 sr@xmath9 ) .
the average @xmath68 versus @xmath167 and best - fit line are shown in figure [ correlation_linear](d ) .
only the data for @xmath153 were used in the fit .
we note a very nice , unexpected , correlation between the fuv background and the diffuse h@xmath2 emission , some of which is known to be emitted from the wim .
as can be seen in figures [ correlation ] and [ correlation_linear ] , the low intensity behavior in the correlation of @xmath68 with @xmath164 is slightly different from in the correlation with either @xmath0 or @xmath163 .
figure [ correlation ] shows that the correlation between @xmath68 and @xmath164 is overall straight down to the lowest intensities , while other relations become flat or curved a bit upward .
moreover , in figure [ correlation_linear ] , at the low intensity of @xmath68 ( @xmath168 cu ) , the diffuse fuv intensity seems to decouple from the 100 @xmath1 m intensity ( and a bit from the column density ) , while the correlation between @xmath68 and @xmath164 is still retained .
these results seem to indicate that the fuv background has a stronger correlation with the h@xmath2 emission , at least qualitatively , than with the other emission tracers , although the correlation coefficient between @xmath68 and @xmath164 is not significantly higher than the other correlation coefficients ( figure [ correlation ] ) .
figure [ correlation_sinb ] shows the results after the removal of inverse-@xmath87 dependence due to the plane - parallel medium , offering clearer evidence that the correlation between @xmath68 and @xmath164 is stronger than other correlations .
since the far - ir emission from interstellar dust correlates well with neutral hydrogen emission ( e.g. , * ? ? ?
* ) , the diffuse fuv continuum correlates well with both @xmath163 and @xmath0 at the same significance , as indicated by the correlation coefficients in figure [ correlation_sinb](a ) and ( b ) .
if the correlation between the fuv and h@xmath2 backgrounds is simply due to the co - existence of dust and h@xmath2-emitting gas , the correlation of @xmath68 and @xmath164 would have the same significance as with either @xmath163 or @xmath0 . on the other hand ,
the correlation with the h@xmath2 background would be enhanced when the physical origin of a large portion of the h@xmath2 background is analogous to the fuv background .
we may then conjecture that the better correlation of the fuv with h@xmath2 intensities is attributed to the similarity of the physical origins of the diffuse fuv and h@xmath2 emissions .
the fact that the correlation between @xmath68 and @xmath164 appears to exist not only on a large scale but also in smaller - cloud scales supports the idea wherein their physical origins are at least in part related .
we discuss this connection briefly in 6.1 and in more detail in a paper in preparation .
we here consider several possible explanations for the relationship between the fuv continuum and h@xmath2 emissions , assuming that the physical origins of the backgrounds are indeed closely related with each other .
first of all , a large fraction of the correlation between @xmath68 and @xmath164 may be caused by the two - photon continuum emission in fuv wavelengths , which originates from the wim .
however , the contribution of two - photon emission to the fuv continuum ( @xmath169% , as found in
[ sub : two - photon ] ) is not large enough to explain the correlation relation eq .
( [ eq : fuv_halpha ] ) .
@xcite estimated that the scattered h@xmath2 intensity at high galactic latitudes is @xmath170% of the total h@xmath2 intensity .
recently , @xcite and @xcite found that in some dust clouds , the h@xmath2 radiation is mostly due to scattering by dust grains that are illuminated by a h@xmath2-emitting source off the line of sight .
@xcite showed that a substantial fraction of the diffuse high - latitude h@xmath2 background is caused by interstellar dust scattering of h@xmath2 photons that originate elsewhere in the galaxy .
in particular , they found many extended regions where the fraction of the scattered h@xmath2 intensity is of an order of 50% or higher .
@xcite and @xcite also found that the most likely average fraction of scattered ha intensity is about 20% of the total ha intensity .
thus , the scattered component of the h@xmath2 emission may strengthen the correlation between the diffuse h@xmath2 and fuv intensities , especially at high latitudes .
it should be noted here that the differences between the data in figure [ correlation_linear](b ) and figure [ correlation_linear](d ) accord with the increasing significance of the dust - scattered h@xmath2 emission with latitude . at low latitudes
the diffuse h@xmath2 emission is mostly from ionized gas and not related to dust - scattering , while the fuv background is from dust - scattering .
therefore , at low latitudes and high values of the diffuse fuv intensity , a much wider range of variability in the h@xmath2 intensity is found for a given value of the diffuse fuv intensity ( figure [ correlation_linear](d ) ) while the diffuse fuv intensity still correlates with 100 @xmath1 m emission showing an enhancement due to the effects of strong forward - scattering ( see 5.1 ) . on the other hand , at high latitudes and low values of the diffuse fuv intensity
, much of the structure in the faint h@xmath2 background may be due to scattering .
thus , we are seeing a linear correlation with the diffuse fuv intensity there .
in other words , the h@xmath2 contribution from dust - scattering becomes increasingly important at higher latitudes .
such a strong linear correlation seems to imply that a significant portion of h@xmath2 at high latitudes is due to dust - scattering .
a more detail comparison of the diffuse fuv and h@xmath2 maps in both large and small scales is needed to address how large portion of the diffuse h@xmath2 emission is due to dust - scattering .
finally , there is one possibility beyond the scattered h@xmath2 that is worth considering .
the ism may have low - density paths and voids that allow for ionizing photons from midplane ob stars to reach and ionize gas many kiloparsecs above the galactic plane @xcite .
pathways that provide lower than average densities to high latitudes could enhance both scattered fuv and gas ionized by the lyman continuum from the same ob stars that are much closer to the galactic plane .
combined with the scattered h@xmath2 emission , the presence of such pathways could provide a stronger correlated link between fuv and h@xmath2 . in this regard
, we note that if ionizing photons could use low - density paths to penetrate a turbulent ism and reach high latitude clouds , such as ldn 1780 , the resulting ionization would be limited to a thin outer shell and the ambient medium of the clouds @xcite .
@xcite estimated that only @xmath220.05 pc thick skin of ldn 1780 would be photoionized .
for such morphology , the optically thick core regions would appear relatively dark , which contradicts the observations described in @xcite and @xcite .
however , we note that this does not preclude ionizing photons travelling to high latitudes .
photoionization would contribute to a portion of the h@xmath2 intensity seen in ldn 1780 , although dust - scattering is the dominant process for the h@xmath2 intensity on the cloud .
therefore , further investigation is needed to disentangle which process ( dust - scattered h@xmath2 emission versus in - situ ionized gas ) is more significant in explaining the correlation between the diffuse fuv and h@xmath2 emissions .
a weak saturation effect in the correlation between @xmath68 and @xmath164 , which is similar to but much weaker than those in the correlation of @xmath68 versus @xmath0 and @xmath152 , is also noticeable .
the weak saturation seen in the correlation with @xmath164 may be easily understood through the fact that the wim can also be approximated with a plane - parallel model but the extinction cross - section at fuv is higher than that of h@xmath2 .
if the diffuse fuv background correlates primarily with galactic latitude @xmath3 , the fuv background intensity is described by @xmath171 .
in addition to these , we also noted in the previous section that the diffuse fuv intensity can be approximated with a simple plane - parallel model defined by @xmath172 .
we thus fitted the data with two trial functions , one with two free parameters and the other with only one free parameter .
the dependence on @xmath92 was also obtained by averaging the diffuse fuv intensity for a given galactic latitude .
best regression functions are then given by @xmath173 and @xmath174 for two - parameters and one - parameter models , respectively .
the average intensity of the fuv background versus @xmath3 are also shown in figure [ correlation_linear ] .
the regression functions with two and one parameters are denoted by solid and dashed lines , respectively .
the value of the two - parameters function at @xmath175 is a bit lower than the value obtained by @xcite , but still within their error range .
we note that the two - parameters model is better in explaining the dependence of @xmath68 on galactic latitude . however , the physical interpretation of the two - parameters model is problematic , as discussed in [ sub : isotropic - component ] .
comparison of the fuv background with the soft x - ray emission has been mentioned by @xcite and @xcite .
especially @xcite claimed there was a good correlation of the fuv intensity in high latitude range ( @xmath176 ) with soft x - ray brightness .
they interpreted the correlation as an indication of the significant contribution of the hot ism to the diffuse fuv background .
however , @xcite argued that hot gas has too small an emission measure to contribute significantly to the fuv background .
soft x - ray emission has been observed extensively with rosat and is well - known to be anti - correlated with column density .
indeed , we found a strong anti - correlation between the diffuse fuv intensity and the soft x - ray background , as shown in figure [ correlation](d ) .
this anti - correlation between the diffuse fuv background and the soft x - ray also supports the dust - scattered origin of the diffuse fuv background .
in presenting the results , we used logarithmic scales to visualize large orders of magnitude in observed fuv intensity . here , we show that the distribution of the diffuse fuv intensity seems to be well represented by a log - normal distribution or , equivalently , a gaussian distribution in a logarithm scale , although its detailed study is beyond the scope of the present paper .
@xcite found that the histogram of the h@xmath2 emission line observed from the wim fits well with a log - normal distribution .
they attributed the log - normal nature to the turbulence in the wim .
a log - normal distribution is the expected probability density function of density and/or column density for the ism with a density structure established by turbulence ( e.g. , * ? ? ?
since the scattered fuv intensity is directly related to the dust column density , the log - normal property of the distribution of the diffuse fuv intensity is a natural consequence of the density structure and the turbulence property of the ism .
thus in figure [ log - normal ] , we plotted intensity histograms of @xmath86 for various galactic latitude ranges together with best - fit log - normal functions , and found that the intensity distributions are indeed well represented with log - normal functions . here , we removed the magellanic clouds and 19 large - scale regions ( the same regions as in @xcite ) with significantly enhanced h@xmath2 emission to avoid discrete ionized regions where the contribution of two - photon emission to @xmath68 may be significant .
we also removed some high intensity `` streaks '' ( most visible in the left panel of figure [ diffuse_map](b ) ) that show up in the exposure maps .
log - normal distribution shows an extended tail to large intensity values in a linear scale .
some evidence of the extended tail in an intensity histogram has already been noticed in the previous studies .
@xcite noted that the distributions of the points in correlations between uv brightnesses and column density are not symmetrical with respect to the linear regression lines .
there was a high flux tail in the distribution of the uv intensities at a given value of @xmath177 ( ) .
the intensity histograms in figure 2 of @xcite also show such extended tails .
these regions with uv intensity excesses were rejected for the correlation studies in @xcite and @xcite by eliminating the points in the tail with counts larger than the median by several times the dispersion and iterating the procedure .
@xcite attributed this property to an excess fuv radiation in certain regions of the sky , perhaps due to two photon emission by ionized gas .
however , @xcite and @xcite showed that the emission from the wim is unlikely to contribute significantly to the general diffuse uv background .
@xcite found that the excess in the intensity histogram does not obey the poisson distribution ; they attributed the high - count skewness and tail to the contribution of the stars brighter than @xmath178 and unresolved faint stars .
however , we argued that contribution of residual stars to the diffuse fuv background is not significant .
in addition , the faint late - type stars can not explain the hardness ratio observed at low intensity regions , as shown in 4 .
@xcite also noted a large variance in the fuv intensity and attributed the observed variance to the `` cosmic variance '' predicted by the multi - cloud model combined with multiple scattering and three - dimensional radiation field variance .
we thus believe that the anomalous high intensity tails shown in previous studies are in fact due to an intrinsic property of the log - normal nature of the diffuse fuv background . a detailed statistical property of the dust density distribution can be investigated through a radiative transfer modeling of the starlight dust - scattered off in turbulent media
. it may also be possible to isolate high - intensity tails in the histograms and identify extended local features in the fuv maps .
some locally enhanced features were found , but bright regions generally had a patchy appearance .
we also found that some of the enhanced features in the fuv sky coincided with the h@xmath2 sky , but not always .
we note here that the observed intensity toward a cloud could be higher , lower , or equal to its surroundings ( as demonstrated in @xcite ) , depending on the scattered intensity , incident intensity from behind the cloud , and cloud optical depth .
therefore , the high - intensity tails in histograms do not necessarily correspond to local dust clouds . in addition , because of the difference in dust - extinction cross - sections at the fuv wavelengths and h@xmath2 line , there would not be a strict correlation between diffuse fuv and h@xmath2 emissions , even when all the fuv and h@xmath2 background photons originate from dust - scattering .
the origin of both the high - intensity tails in the histograms and the diffuse h@xmath2 line emission may be elucidated by more detailed analyses of the dust clouds observed in both the fuv wavelengths and h@xmath2 line , along with self - consistent radiative transfer models .
but , such analyses are beyond the scope of the present paper .
we found a weak reddening of the diffuse radiation relative to the direct starlight in 4 . in order to understand the trend , we calculated the scattered to direct intensity ratio for a simplified case where a point source is embedded in a homogeneous spherical dust cloud .
as noted in @xcite , the analytical expression derived by @xcite approximates the scattered to the unattenuated luminosity ratio very well . in the approximation ,
the scattered intensity is given by @xmath179 where @xmath180 and @xmath181 , and the attenuated direct starlight by @xmath182 . here
, the quantities @xmath183 , @xmath43 , @xmath184 , and @xmath185 refer to the unattenuated stellar luminosity , grain albedo , phase function asymmetry factor , and radial optical depth of the spherical cloud measured from the center to the outer edge .
we calculated ratios @xmath186 , @xmath187 , and @xmath188 as a function of wavelength using the extinction cross - section @xmath189 , @xmath190 , and @xmath191 from @xcite for an illustrative case of @xmath192 @xmath193 in figure [ aeffect ] . here ,
@xmath194 , @xmath195 , and @xmath196 are the input , direct , and scattered luminosities .
note that the albedo @xmath190 is an increasing function with the wavelength .
for comparison , we also plotted the results calculated for three hypothetical cases : ( 1 ) the case of a constant albedo of @xmath197 ) and a variable asymmetry factor @xmath191 , ( 2 ) the case where both @xmath43 and @xmath184 are constants obtained at 1550 , and ( 3 ) the case in which the wavelength - dependence of @xmath190 is reversed so that the albedo decreases with the wavelength .
the figure shows that assuming the asymmetry factor to be a constant does not significantly affect the result while the albedo does .
as is evident , the directly - escaped spectrum is redder or softer than the input spectrum ( figure [ aeffect]a ) .
when a constant albedo is assumed , we obtain a scattered spectrum reddened compared to the unattenuated input spectrum , but bluer than the ( attenuated ) directly - escaped spectrum because of the rapid increase of the absorption with the wavelength . however , as the albedo increases with the wavelength in the model of @xcite and @xcite , the scattered spectra rise much more rapidly with the wavelength than the attenuation , thereby giving a relatively - red scattered spectrum compared to the directly escaped spectrum . in the case in which the wavelength - dependence of @xmath190 is reversed ,
the scattered spectrum become bluer relative to both the unattenuated input and the reddened directly - escaped spectra , which is contrary to the other cases . the same trend is observed when the optical depth is varied . we therefore conclude that the relative softness of the diffuse background radiation compared to the stellar radiation field is attributed to the increase of dust albedo with the wavelength . in 4 , we noticed a weak rise in the diffuse spectra longward of 1550 .
the rise itself tells nothing about the wavelength - dependence of dust albedo unless the incident unattenuated spectrum @xmath198 is known . instead
, the relative softness or hardness of the scattered radiation compared to the directly observed radiation can give direct information about the wavelength - dependence of dust albedo .
in this regard , we note that the dust albedo provided by @xcite is constant or decreases with the wavelength in fuv , and thus is not consistent with our results .
a more detailed analysis on the spectral shape can be obtained by radiative transfer models including wavelength dependence in calculating stellar radiation fields rather than that performed for a single wavelength band as have been done by @xcite and @xcite .
the wim , which is expected to have a temperature of @xmath199 k , must be a source of two - photon continuum in the fuv wavelengths , potentially impacting the interpretation of the fuv continuum background @xcite .
the two - photon continuum intensity at @xmath22 1600 due to the wim with a temperature of 8000 k is estimated to be @xmath200 cu ( @xcite ; see also figure [ two_photon ] ) .
using the average h@xmath2 intensity distribution @xmath201 r , which was obtained from rather limited observations , reynolds also obtained @xmath202 cu .
however , using the result of @xcite obtained from the full wim observations , the distribution of h@xmath2 for @xmath203 is fitted on average by @xmath204 r ( see also @xcite ) .
this leads to a lower two - photon intensity of @xmath205 cu , which accounts for @xmath206% of the fuv background ( eq . [ eq : fuv_sinb2 ] ) .
the two - photon continuum s contribution to the diffuse fuv intensity is reduced even further when the model of wim developed by @xcite is adopted .
recent observations using the _ wilkinson microwave anisotropy probe ( wmap ) _ have found that the ratio of the free - free radio continuum to h@xmath2 is surprisingly low in the wim @xcite .
@xcite considered a three - component model consisting of a mix of ( 1 ) photoionized hot gas , ( 2 ) gas that is recombining and cooling after removal of a photoionizing source , and ( 3 ) cool gas . in their standard model for explaining the observed intensity ratios of the free - free radio continuum to h@xmath2 and [ @xmath366583 to h@xmath2 , the authors assumed that the scattered fraction of the h@xmath2 originating from the hot gas is @xmath207 , while the fractions of the hot , cooling , and gases are @xmath208 , @xmath209 , and @xmath210 , respectively .
the temperature of the hot gas was found to be @xmath211 k. the cooling gas cooled down rapidly from its initial temperature @xmath212 10,000 k to @xmath213 k ( fig .
3 of @xcite ) . using the two - photon spectral profile @xcite ,
the hydrogen recombination coefficient @xmath214 for case b ( table 3 of @xcite ) , and the h@xmath2 emission rate ( eq.10 of @xcite ) , we estimated the two - photon intensity averaged over 13701710 from hot and cooling ionized gas ( figure [ two_photon ] ) : @xmath215 cu , and @xmath216 cu .
the two - photon intensity of the cooling gas is mostly determined by the gas with a temperature of 400 k , because the cooling gas is 400 k or lower for more than 98% of its lifetime ( fig .
3 of @xcite ) .
the total two - photon intensity originating from the ionized gas is then given by @xmath217 cu .
the distribution of two - photon emission is then @xmath218 cu , corresponding to @xmath219% of the fuv background intensity . in summary , two - photon emission may account for @xmath126% of the diffuse fuv background . although the two - photon emission may be significant in discrete bright regions , the contribution of two - photon emission to the diffuse fuv background is much smaller than has previously been suggested .
the large variation of @xmath68 ( from a factor of 23 up to an order of magnitude ) is also unlikely to be the result of two - photon emission .
the phrase `` isotropic component '' has been used to refer to the diffuse intensity extrapolated to @xmath220 , or , alternatively , to @xmath221 , and is frequently assumed to be extragalactic background .
@xcite claimed that the isotropic extragalactic component of the diffuse fuv background should be obtained by extrapolating to @xmath222 , instead of to @xmath223 as practiced by @xcite .
@xcite discussed systematic errors arising in the method of extrapolating high latitude measurements of the diffuse fuv background to @xmath220 or to @xmath222 to derive the level of the cosmic extragalactic background .
they argue that the derivation of the cosmic background from the extrapolation of the diffuse fuv background to @xmath220 is more likely to yield a reliable result .
they found , in fact , that extrapolations of the predicted fuv background intensities through a monte - carlo radiative transfer model to @xmath224 produce negative values when it should produce a zero intercept .
we also found from the spear data that the extrapolation to @xmath224 indeed produces a negative value of @xmath225 cu .
a discussion of the origin of the isotropic component is beyond the scope of the present paper .
@xcite observed intensities as low as 300 cu at high galactic latitudes .
@xcite placed constraints on the extragalactic background @xmath226 cu .
we also found the isotropic component consistent with previous authors , but a bit higher .
it should be noted that a similar level of the isotropic component was found from the relationship between the diffuse fuv and h@xmath2 emissions ( eq . [ eq : fuv_halpha ] ) .
however , it is not clear at this time that all of the @xmath134 cu , the extrapolation value of the average fuv-@xmath0 correlation to @xmath220 , corresponds to the extragalactic background .
in fact , we observed the fuv intensities lower than 300 cu as is obvious in figure [ histograms ] .
@xcite analyzed the fuv data observed in the faust experiment @xcite and estimated a contribution of about 700@xmath227200 cu that is uncorrelated with galactic parameters .
they argued that the extragalactic component of more than 300 cu is excluded by the lower limit directly measured by @xcite and concluded that most likely value of isotropic extragalactic background radiation would be 160@xmath22750 cu , after correction for galactic extinction .
in this paper , we reported some general results on the diffuse fuv background and the total fuv radiation observed with spear / fims .
our main conclusions are as follows : 1 .
strong anisotropies are found in not only the diffuse background intensity but also in the hardness ratio ( 13701520 to 15601710 ) .
2 . we found a good correlation between the diffuse fuv background and other galactic quantities , such as column density , 100 @xmath1 m emission , and h@xmath2 emission .
the correlation of the fuv continuum background with the diffuse h@xmath2 emission seems to be better than with the other quantities .
anti - correlation with the soft x - ray was also found .
correlation of the fuv background with the direct stellar intensity is rather weaker than the correlation with other waveband observations , such as 100 @xmath1 m and h@xmath2 emissions .
3 . the fact that a linear correlation of the fuv background with the diffuse h@xmath2 emission is found at low intensities seems to indicate that the dust - scattered component of h@xmath2 becomes increasingly important at higher latitudes .
4 . the spatially - averaged total fuv intensity , including the direct stellar and diffuse fuv intensities , observed with spear / fims is weaker than the well - known isrf models , such as the habing , mathis , and draine models .
the spectrum of the diffuse fuv background is in general flat and a bit softer than the directly - escaped stellar spectrum .
the relative softening can be attributed to the rise of dust albedo as the wavelength increases .
the hardness ratio seems to follow in general the longitudinal distribution of ob - type stars in the galactic plane .
the `` isotropic component '' obtained by comparing the fuv background with other ism tracers , @xmath0 , 100 @xmath1 m , and h@xmath2 emissions is consistent with the previous results .
evidence that the intensity histogram of the diffuse background is well represented by log - normal distribution is found , as for the diffuse h@xmath2 emission .
the above results well accord with the fact that most of the diffuse fuv background is scattered light of the fuv stellar radiation by the interstellar dust . @xcite and @xcite developed radiative transfer models to interpret the diffuse fuv background .
our future work will deal with a monte - carlo simulation model incorporating three - dimensional stellar distribution obtained from the _ hipparcos _ catalog and three - dimensional dust distributions modeled by several authors such as @xcite and @xcite .
the model compared with the _ spear / fims _ data should provide a better understanding to the observational results .
the spear / fims was supported by nasa grant nag5 - 5355 and flew on the stsat-1 mission , supported of the korea ministry of science and technology .
we acknowledge the use of the legacy archive for microwave background data analysis ( lambda ) .
support for lambda is provided by the nasa office of space science .
s. was supported by a national research foundation of korea grant funded by the korean government ( grant no .
313 - 2008 - 2-c00377 ) .
we thank the anonymous referee for comments that led to improvements in the manuscript .
pixel by pixel comparison of the 100% and 10% shutter aperture data . by fitting the data with a linear function , the scale factor for the 10% shutter aperture data and detector background rate are estimated .
the comparison is made with the @xmath25 maps .
here , cu represents the continuum unit ( photons @xmath8 s@xmath9 @xmath9 sr@xmath9 cu ) .
] instrumental background spectrum .
the spectrum is essentially the inverse of the effective area curves , as shown in @xcite .
an @xmath62 error range for the average of detector background value is also shown with dotted lines . ]
histograms of ( a ) exposure time ( s ) , ( b ) the fuv intensity ( photons s@xmath9 cm@xmath9 sr@xmath9 @xmath9 , cu ) over the observed sky , and ( c ) signal - to - noise ratio with bins scaled logarithmically .
the histograms are made with about @xmath229 pixels ( @xmath25 ) .
a background of 190 cu was removed from the data to take into account the instrumental dark background . ] the diffuse fuv intensity versus galactic longitude .
the top panel plots the two - dimensional histogram of the diffuse fuv intensity as a function of galactic longitude .
the bottom panel plots the median value within each of the @xmath148 longitude strips from the data .
the vertical extent of each plotted bar is determined from the average deviation about the median within that bin . ]
the diffuse fuv intensity versus galactic latitude .
the top panel plots the two - dimensional histogram of the diffuse fuv intensity as a function of galactic latitude .
the bottom panel plots the median value within each of the @xmath230 latitude strips from the data .
the vertical extent of each plotted bar is determined from the average deviation about the median within that bin . ]
@xmath86 versus @xmath87 .
the top panel plots the two - dimensional histogram as a function of galactic longitude .
the bottom panel plots the median value within each of the @xmath231 latitude strips from the data .
the vertical extent of each plotted bar is determined from the average deviation about the median within that bin .
the black and red points represent the data observed toward northern and southern galactic latitude , respectively .
the horizontal line denotes the median @xmath232 value . ]
spear / fims spectra averaged over all the observations are compared with the three models of the isrf .
the four models are from @xcite , @xcite , @xcite , and @xcite . the spectrum measured by @xcite
is also compared . ]
correlation between the average hardness ratio ( 13701520 to 15601710 ) and the number of ob - type stars .
the average hardness ratio and the number of ob - type stars were calculated within each of the @xmath148 longitude strips , and then the average hardness ratio versus the number of ob - type stars was plotted .
the emission line was excluded in estimating the hardness ratio . ]
scattering and absorption in a homogeneous dust medium .
( a ) the directly - escaped and ( b ) scattered spectra relative to the input luminosity of the point source .
( c ) the scattered spectrum relative to the attenuated direct stellar spectrum . here , various combinations of the wavelength - dependences of the dust albedo and asymmetry factor are examined . ]
two - photon continuum intensity that accompany 1 r of h@xmath2 emission from ionized interstellar hydrogen .
the continuum intensity is shown for three temperatures : 9100 k , solid line ; 8000 k , dashed line ; 400 k , dotted line . ] | we present the general properties of the far - ultraviolet ( fuv ; 13701710 ) continuum background over most of the sky , obtained with the _ spectroscopy of plasma evolution from astrophysical radiation _ instrument ( spear , also known as fims ) , flown aboard the stsat-1 satellite mission .
we find that the diffuse fuv continuum intensity is well correlated with @xmath0 , 100 @xmath1 m , and h@xmath2 intensities but anti - correlated with soft x - ray intensity .
the correlation of the diffuse background with the direct stellar flux is weaker than the correlation with other parameters .
the continuum spectra are relatively flat .
however , a weak softening of the fuv spectra toward some sight lines , mostly at high galactic latitudes , is found not only in direct - stellar but also in diffuse background spectra .
the diffuse background is relatively softer than the direct stellar spectrum .
we also find that the diffuse fuv background averaged over the sky has a bit softer spectrum compared to direct stellar radiation .
a map of the ratio of 13701520 to 15601710 band intensity shows that the sky is divided into roughly two parts .
however , this map shows a lot of patchy structures on small scales .
the spatial variation of the hardness ratio seems to be largely determined by the longitudinal distribution of ob - type stars in the galactic plane .
a correlation of the hardness ratio with the fuv intensity at high intensities is found but an anti - correlation at low intensities .
we also find evidence that the fuv intensity distribution is log - normal in nature . |
the increasing importance of coherent phonon spectroscopy in material science is related to the growing problem of heat dissipation in modern nanoscale devices .
this problem is impossible to solve without detailed understanding of underlaying phonon - phonon and phonon - electron interactions on the nanoscale .
one of the methods to study this processes is coherent phonon spectroscopy , in which a particular phonon spectrum is excited coherently in the sample and detected optically .
research efforts in this direction resulted in significant progress in generation and detection of coherent phonons in various materials .
the available phonon frequency has reached the thz acoustic limit @xcite and basically the whole phonon frequency range nowadays could be excited coherently . however , convenient optical detection methods based on raman@xcite or brillouin@xcite scattering allow for the observation of phonons excited only in the vicinity of the brillouin zone center .
therefore , sub - thz acoustic phonons could be accessed optically only in multilayer structures , in which the acoustic dispersion branch backfolds many times inside a mini - brillouin zone of a multilayer @xcite .
modern progress in pulsed laser techniques as well as in multilayer fabrication has lead to a set of successful experiments in which the coherent zone - folded superlattice phonons have been optically excited and detected @xcite .
however , these optical methods are insensitive to the thz frequency phonons which have propagated into the bulk of the crystal due to the unfolding of the phonon dispersion curve .
convenient optical methods based on brillouin scattering in this case have a detection limit in the 100ghz range given by the wave vector magnitude of the optical light @xcite . recently ,
ultrafast x - ray diffraction ( uxrd ) has become available to extend the accessible phonon frequency range to above 100ghz .
it has been used successfully to study both the time - domain structure of optically excited zone - folded coherent acoustic phonons in epitaxial multilayers@xcite as well as to observe the propagation of unfolded phonons into the bulk @xcite .
in this paper we report our new uxrd experiments from coherent quasi - monochromatic longitudinal acoustic phonons in srtio@xmath0 synthesized by fs - laser excitation of srruo@xmath0/srtio@xmath0 ( sro / sto ) epitaxial multilayers .
using uxrd we determine the laser excited phonon spectrum in srtio@xmath0 and monitor the modification of the spectrum in the time domain .
the epitaxial multilayers were prepared using pulsed laser deposition @xcite .
the experiments are carried out at the bessy edr beamline using a unique setup for a 1mhz repetition rate uxrd experiments .
the experiments are done in a traditional scheme which uses an optical delay line to change the time interval between the optical pump and the x - ray probe pulses .
we use infrared optical pulses with the wavelength of 1.03@xmath3 m for pumping and 8kev x - rays for probing the lattice dynamics .
the important feature of this setup is the simultaneous acquisition of the x - ray photons scattered from the sample before and after the pumping optical pulse .
this makes the x - ray intensity difference signal sensitive only to those changes in the crystal lattice which were exclusively initiated by the optical pulses . for a further details we refer to a recent publication describing the setup@xcite
recent studies showed that the optical excitation of a metal transducer by a sequence of ultrashort laser pulses is an efficient method to generate sub - thz quasi - monochromatic longitudinal acoustic ( la ) phonons @xcite .
in essence , the repetitive generation of bipolar strain pulses by the laser - excited transducer @xcite forms a phonon wavepacket of narrow spectral bandwidth propagating throughout the substrate . in this report
we consider a different approach which uses a spatial repetition instead of a temporal one , i.e. , the excitation of a periodic metal - dielectric multilayer ( superlattice ) with a single ultrashort laser pulse ( see fig . [ phononsimulations](a ) ) . /srtio@xmath0
multilayer on top of the srtio@xmath0 substrate .
( b ) calculated instant strain profile in the sample along the surface normal , taken at 250ps after the multilayer excitation .
( c ) calculated spectrum of the strain pulse in the wave vector and frequency domains induced by the optical excitation .
the vertical scale stands for the excitation fluence of 6mj/@xmath4 used in the experiment . ]
this way the so - called superlattice phonon mode is excited @xcite which subsequently unfolds into the substrate thereby forming la phonon wavepackets with similarly narrow spectral bandwidth @xcite . in most cases the laser - induced coherent lattice dynamics may be calculated using either a model of a continuous elastic medium @xcite or a linear - chain model ( lcm ) of masses and springs @xcite . here
we employ the latter approach which will be more appropriate for short - period superlattices and automatically accounts for the acoustic phonon dispersion .
the details of the numerical model can be found in ref . .
the excellent agreement of such a linear - chain model with related optical and uxrd experiments has been previously demonstrated for various sample structures @xcite .
further we consider linear chain calculations for a five - period sro / sto multilayer with the spatial period @xmath5 nm which is schematically shown in fig . [ phononsimulations](a ) .
the structural parameters of the sample which are determined by static high - resolution x - ray diffraction ( hrxrd ) @xcite are collected in table [ sampleparameterstable ] .
in addition , the longitudinal sound velocities of the individual materials are shown@xcite .
the results of the linear - chain calculations using the parameters given in table [ sampleparameterstable ] and the experimental pump fluence of 6 mj/@xmath4 are shown in the figs .
[ phononsimulations](b ) and ( c ) .
the graph in fig .
[ phononsimulations](b ) shows the calculated one - dimensional strain profile in the sample 250 ps after the excitation .
figure [ phononsimulations](c ) plots the spectral amplitude of the linear - chain eigenvectors ( normal modes ) as a function of the eigenfrequency .
this amplitude spectrum is solely determined by the initial conditions @xcite . for the bulk sto substrate
the eigenvectors are plane elastic waves with wave vector @xmath2 satisfying the well - known dispersion relation of acoustic phonons @xcite . as fig .
[ phononsimulations](b ) illustrates , the optical excitation of the metal layers of the multilayer system results in the generation of a coherent strain wavepacket propagating into the sto substrate at the longitudinal sound velocity @xcite .
the resulting wavepacket inside the sto substrate attains the particular shape shown in fig .
[ phononsimulations](b ) , namely , five leading compression pulses and five trailing expansion pulses which are separated by @xmath6 nm , respectively . in other words ,
the metal / dielectric multilayer acts as the photoacoustic transducer synthesizing the coherent la phonon wavepacket in the sto substrate .
the sharp static profile of the thermal strain inside the multilayer ( see fig . [ phononsimulations](b ) ) remains unchanged with time because the linear chain model neglects the effect of heat diffusion .
the heat diffusion in multilayers is a complicated separate topic which lays out of the scope of this article . in this article we focus on the coherent lattice dynamics in the sto substrate which occur at a later timescale when the effect of heat diffusion within the multilayer does not play a role . for the detailed description of the wavepacket strain profile and its generation we refer to our earlier works @xcite .
.structural and mechanical properties employed in the calculations .
[ cols="<,<,<,<",options="header " , ] [ table ] there are two major reasons for the observed decrease of the uxrd peak intensities . first , the absorption of the x - rays by the crystal reduce the sensitivity of the x - rays to the wavepacket as it propagates deeper into the substrate .
second , the dissipation of energy from the elastic wave due to the finite phonon lifetimes leads to a decay of the strain amplitude . considering the first reason , the decay time of the uxrd signal exclusively due to x - ray absorption is related to the x - ray absorption coefficient @xmath7 by @xmath8 where @xmath9 is the longitudinal sound speed in the substrate@xcite .
the relative variation of the x - ray scattering vector @xmath10 during the presented uxrd experiments is @xmath11 which implies that @xmath12 is virtually independent of the observed phonon wavevectors @xmath2 . in the measured off - bragg region
the x - ray extinction due to dynamical x - ray diffraction is negligible , therefore only the angular independent x - ray absorption is relevant . under the chosen experimental conditions we estimate a signal decay time of @xmath13 ps due to the x - ray absorption . nevertheless , since this value is critical for the correct interpretation of the experimental data , we have performed dynamical xrd calculations based on results of the linear - chain lattice dynamics in harmonic approximation which excludes the effect of phonon damping@xcite .
the simulations yield the @xmath2-independent value of 450@xmath145ps for the decay constant due to x - ray absorption which we will use in the following . regarding the second reason for the decay of the uxrd phonon signals , we assume an exponential law for the decrease of the phonon population @xmath15 and define the associated decay time @xmath16 . according to ( [ amplitudefromintensity ] ) the corresponding intensity of the scattered x - rays possesses the same decay constant .
therefore , the uxrd signal decay mechanisms introduce the relationship between the phonon lifetimes @xmath16 , x - ray absorption time constant @xmath12 and the experimentally measured time constant @xmath17 : @xmath18 the phonon lifetimes @xmath16 extracted from the uxrd experiments according to ( [ decaytimesrelationship ] ) are given in the table ( [ table ] ) for the wave vector magnitudes corresponding to the first and second order phonon peaks .
the standard deviation @xmath19 for the phonon lifetimes are calculated from the standard deviations @xmath20 of the experimental time constants according to the error propagation relation .
the important observation is the fact that the determined lifetimes differ by a factor of approx .
4 whereas the related phonon wave vector differ by a factor of 2 .
in addition to the signal decay we see a gradual drift of the phonon peak position to higher values of @xmath2 as the wavepacket propagates .
this is seen from the fig .
[ phononpeakshift](a ) in which the phonon spectrum is shown for different time delays . in the following section we discuss the physical interpretation of the described observations .
the central observation from the transient phonon - induced diffraction peaks presented in the previous section is the apparent quadratic decrease of the phonon lifetime with phonon wave vector @xmath2 .
this observation is in agreement with the @xmath21 law predicted by the akhiezer s sound attenuation mechanism @xcite .
there are basically two theories which explain the attenuation of a hypersonic waves in dielectric crystals due to incoherent anharmonic phonon - phonon scattering .
these are the landau - rumer theory @xcite and the akhiezer theory @xcite which have different application limits : @xmath22 where @xmath23 is the angular frequency of the hypersonic wave and @xmath24 is the mean thermal phonon relaxation time .
that is , in both cases the elastic energy of hypersound waves decreases with time due to the interaction with incoherent thermal phonons .
the mean thermal phonon relaxation time @xmath24 can be estimated from the thermal conductivity @xmath25 , heat capacity @xmath26 , average sound speed @xmath27 and mass density @xmath28 by the relation @xcite @xmath29 for sto at room temperature ( [ thermalconduct ] ) yields @xmath30 ps . the experiments are performed at room temperature .
more precisely , the sample persisted at around 400k during the actual measurements due to the thermal load from the laser@xcite . condition ( [ akhiezer ] ) is fulfilled for this temperature range .
that is , for the presented experiments akhiezer s theory of relaxation damping could be applied , hereby explaining the observed ratio of the phonon lifetimes for the first and second - order phonon diffraction peaks .
however the uxrd data exhibit additional features which can not be explained within akhiezer s sound attenuation model .
we extracted the transient peak positions by gaussian fits to the data shown in fig .
[ phononpeakshift](a ) and the results are plotted as symbols in fig .
[ phononpeakshift](b ) .
the solid lines indicate linear fits to the phonon peak positions as a function of time .
clearly , a gradual shift of the first and second - order spectral components to higher @xmath2-values can be observed as the phonon wavepacket propagates deeper into the sto substrate .
a recent study revealed the nonlinear propagation of large - amplitude sound wavepackets in sto at room temperature @xcite .
in addition to the akhiezer - like attenuation of the coherent la phonons the authors observed transient changes of the acoustic spectrum due to coherent anharmonic phonon - phonon scattering within the wavepacket .
the lattice anharmonicity gave rise to a strain - dependent longitudinal sound velocity .
in particular , the sound velocity of compressive ( tensile ) parts of the wavepacket was found to increase ( decrease ) with the strain amplitude .
this effect led to an anomalous dispersion of the wavepacket and the corresponding modification the phonon spectrum .
accordingly , we expect the first compressive half of the wavepacket shown in fig .
[ phononsimulations](b ) to propagate faster than the second tensile half .
moreover , the individual pulses inside the respective parts also exhibit different velocities due to the exponential amplitude distribution determined by the optical penetration of the pump light in sro .
for both the compressive and tensile parts of the strain pulse the spatial separation @xmath31 of the individual pulses of the wavepacket is reduced as it propagates , i. e. the wavelength of both sub - packets is decreased .
this explains the observed shift of the phonon peaks to larger @xmath2-values . the presented uxrd data thus evidence the influence of two different effects on the propagation of la phonon wavepackets generated by periodic multilayers
first , the inevitable attenuation of the wavepackets by akhiezer s relaxation damping and , second , the change of spatial and spectral shape of the wavepacket by nonlinear sound propagation .
both effects influence the observed phonon lifetime , however , for a quantitative determination of the respective contributions additional measurements have to be performed .
the results of our earlier all - optical experiments having much stronger excitation were successfully explained solely in the framework of nonlinear acoustics@xcite .
we believe that at the presented experimental conditions the influences of both damping mechanisms the non - linear acoustic propagation and the akhiezer s relaxation are comparable . uxrd has the advantage of measuring the lattice dynamics directly and quantitatively , i.e. the absolute amplitude of the lattice motion is determined .
the wavevector range over which acoustic phonons in bulk material are accessible is very large . in particular resolving the second order phonon peak as presented in this paper or higher orders
is possible .
the extension of the uxrd detection of acoustic phonons in amorphous materials is challenging the available x - ray fluence , since the bragg spots are dispersed in diffraction rings . on the other hand
all - optical picosecond acoustics @xcite , in principle , do not require crystalline materials and for transparent media , the propagation of strain pulses can be monitored over longer distances . with current technology femtosecond time resolution
is standard in all - optical experiments , while it is still a challenge in x - ray technology , which essentially resolved by free - electron lasers .
high time - resolution permits the determination of the wave vector dependent sound velocity in addition to the damping time .
we believe that the uxrd - based methods and the all - optical methods do not compete with each other but complete each other , together providing a more complete picture of the complex coherent phonon dynamics for a broader range of frequencies and wavevectors and for a broader class of materials and experimental conditions .
this report presents ultrafast x - ray diffraction ( uxrd ) studies on laser - excited periodic srruo@xmath0/srtio@xmath0 multilayers which are epitaxially grown on a srtio@xmath0 substrate .
the ultrafast heating of the metallic srruo@xmath0 layers by ultrashort laser pulses generates coherent longitudinal acoustic phonons which eventually propagate into the substrate as a quasi - monochromatic coherent la phonon wavepacket at hypersonic frequencies .
we discussed the properties of such wavepackets in detail and derive equations which show that uxrd is a powerful tool to measure the spectral phonon population and its dynamics . the presented uxrd data evidence the formation of quasi - monochromatic coherent phonon wavepacket .
we extract the phonon lifetimes of the first and second - order peaks of the phonon spectrum .
the observed quadratic decrease of the phonon lifetime with increasing phonon wavevector @xmath2 is in accordance with akhiezer s mechanism of relaxation damping .
shifts of the peaks corresponding to the excited phonons to larger q - values are interpreted as a modification of the spatial shape profile due to the non - linear wave propagation leading to a strain dependent sound velocity .
this considerably modifies the observed phonon lifetimes .
in essence , uxrd provides a detailed and direct view on the complex nonlinear evolution of phonon - wavepackets , including incoherent damping of the phonon amplitude by coupling to other modes and specific coherent changes of the wavevector spectrum .
we thank the bmbf for funding via 05k 2012-oxide .
39ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1063/1.2750401 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1103/physrevlett.65.764 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevlett.45.298 [ * * , ( ) ] link:\doibase 10.1103/physrevb.31.2080 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.82.1044 [ * * , ( ) ] link:\doibase 10.1063/1.3593488 [ * * , ( ) ] @noop ( ) link:\doibase 10.1126/science.1104739 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.101.025505 [ * * , ( ) ] link:\doibase 10.1002/adfm.200800560 [ * * , ( ) ] link:\doibase 10.1063/1.4727872 [ * * , ( ) ] link:\doibase 10.1063/1.2032589 [ * * , ( ) ] link:\doibase 10.1063/1.3595275 [ * * , ( ) ] link:\doibase 10.1063/1.3688492 [ * * , ( ) ] link:\doibase 10.1103/physrevb.34.4129 [ * * , ( ) ] @noop ( ) link:\doibase 10.1063/1.3402773 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/14/1/013004 [ * * , ( ) ] link:\doibase 10.1007/s00339 - 011 - 6719-z [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.224302 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.144306 [ * * , ( ) ] @noop ( ) \doibase http://dx.doi.org/10.1088/1367-2630/13/9/093032 [ * * , ( ) ] link:\doibase 10.1103/physrev.129.90 [ * * , ( ) ] @noop _ _ , ed .
( , , ) @noop * * , ( ) link:\doibase 10.1016/s0921 - 4526(98)00265 - 8 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1107/s0021889897008807 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.84.111 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.86.3072 [ * * , ( ) ] link:\doibase 10.1007/s003390201421 [ * * , ( ) ] , @noop _ _ , ed .
( , , ) @noop ( ) @noop * * , ( ) @noop * * , ( ) | we synthesize sub - thz longitudinal quasi - monochromatic acoustic phonons in a srtio@xmath0 single crystal using a srruo@xmath0/srtio@xmath0 superlattice as an optical - acoustic transducer .
the generated acoustic phonon spectrum is determined using ultrafast x - ray diffraction .
the analysis of the generated phonon spectrum in the time domain reveals a k - vector dependent phonon lifetime .
it is observed that even at sub - thz frequencies the phonon lifetime agrees with the 1/@xmath1 power law known from akhiezer s model for hyper sound attenuation . the observed shift of the synthesized spectrum to the higher @xmath2 is discussed in the framework of non - linear effects appearing due to the high amplitude of the synthesized phonons .
= 1 |
started to work , the two @xmath1factories open an excited new era in studying @xmath1 meson decays @xcite .
the main research program of these factories is studying cp violation in @xmath1 meson system and investigating their decays . from theoretical point of view
, interest to the rare decays can be attributed to the fact that they occur at loop level in the standard model ( sm ) and they are very sensitive to the flavor structure of the sm as well as to the new physics beyond the sm . from experimental point of view studying radiative
@xmath1 meson decays can provide us essential information on the parameters of the sm , such as the elements of the cabibbo
maskawa ( ckm ) matrix , the leptonic decay constants etc .
, which are yet poorly known .
it is well known that the flavor changing neutral current process @xmath2 has helicity suppression .
these decays are proportional to the lepton mass and because of this reason the decay width of these processes are too small to be measured for the light lepton modes .
it should be noted that in the sm the branching ratio of the @xmath3 and @xmath4 , respectively .
although @xmath5 channel is free of this suppression , its experimental detection is quite hard due to the low efficiency .
it has been observed that the radiative leptonic @xmath6 decays have larger branching ratio compared to that of the purely leptonic models @xcite@xcite .
it was shown in @xcite that similar situation takes place for the radiative decays @xmath7 . in these
decays the contribution of the diagram when photon is radiated from an intermediate charged line , can be neglected , since it is strongly suppressed by a factor @xmath8 .
moreover , the internal bremsstrahlung ( ib ) part when photon is emitted from external charged leptons is proportional to lepton mass , which follows from helicity arguments , gives small contribution .
for this reason in @xmath9 decay the main contribution should come from the diagrams , when photon is emitted from the initial quarks , i.e. , structure dependent part ( sd ) , since they are free of the helicity suppression .
therefore the decay rate of the @xmath10 might have an enhancement in comparison to the pure leptonic models of @xmath2 decay if the sd contributions to the decays are dominant and hence @xmath11 decay might be sensitive to the new physics effects beyond sm .
new physics effects in rare @xmath12 decays can appear in two different ways ; either through new contributions to the wilson coefficients existing in the sm or through the new operators in the effective hamiltonian which are absent in the sm .
the goal of this work is combining both these approaches to study the sensitivity of of the physically measurable quantities , like branching ratio , photon energy distribution , to the new physics effects .
the work is organized as follows . in section 2 , we derive the general expression for the photon energy distribution using the most general form of four fermi interaction . in section 3
we investigate the sensitivity of photon energy distribution and branching ratio to the new wilson coefficients .
in this section we calculate the photon energy distribution and branching ratio for the @xmath11 decay using the most general model independent form of the effective hamiltonian .
the matrix element for the process @xmath9 can be obtained from that of the purely leptonic @xmath13 decay .
the effective @xmath14 transition can be written in terms of twelve model independent four fermi interactions can be written in the following form @xcite : [ effh ] h_eff & = & v_tqv_tb^\ { c_sl |q i _ l b |^+ c_br |q i _ r b |^ + & & + c_ll^tot |q_l _ b_l |_l ^_l + c_lr^tot |q_l _ b_l |_r ^_r + c_rl |q_r _ b_r |_l ^_l + & & + c_rr |q_r _ b_r |_r ^_r + c_lrlr |q_l b_r |_l _ r + c_rllr |q_r b_l |_l _
r + & & + c_lrrl |q_l b_r |_r _
l + c_rlrl |q_r b_l |_r _
l+ c_t |q _ b |^ + & & + i c_te^ |q _ b | _ } , where the chiral projection operators @xmath15 and @xmath16 in ( [ effh ] ) are defined as l = , r = , and @xmath17 are the coefficients of the four fermi interactions .
it can easily be seen from eq .
( [ effh ] ) that several of all wilson coefficients do already exist in the sm .
the coefficients @xmath18 and @xmath19 correspond to @xmath20 and @xmath21 in the sm , respectively .
the next four terms in this expression are the vector interactions .
the interaction terms containing @xmath22 and @xmath23 exist in the sm in the form @xmath24 and @xmath25 , respectively .
therefore @xmath22 and @xmath23 describe the contributions coming from the sm and the new physics , whose explicit are c_ll^tot & = & c_9^eff - c_10 + c_ll , + c_lr^tot & = & c_9^eff + c_10 + c_lr .
the terms with coefficients @xmath26 , @xmath27 , @xmath28 and @xmath29 describe the scalar type interactions .
the last two terms in eq .
( [ effh ] ) with the coefficients @xmath30 and @xmath31 describe the tensor type interactions . having presented the general form of the effective hamiltonian the next problem is calculation of the matrix element of the @xmath32 decay .
this matrix element can be written as the sum of the structure dependent and internal bremsstrahlung parts = m_sd+m_ib .
it follows from eq .
( [ effh ] ) that , in order to calculate the matrix element @xmath33 for the structure dependent part , the following matrix elements are needed [ mel ] & & |s _ ( 1 _ 5 ) b b , + & & |s _ b b , + & & |s ( 1 _ 5 ) b b .
the first two of the matrix elements in eq .
( [ mel ] ) are defined as @xcite [ mel1 ] ( k ) |q _ ( 1 _ 5 ) b b(p_b ) & = & \ { _ ^ q^k^g(q^2 ) + & & i f(q^2 ) } , + + [ mel2 ] ( k ) |q _ b b(p_b ) & = & _ , respectively , where @xmath34 and @xmath35 are the four vector polarization and four momentum of the photon , respectively , @xmath36 is the momentum transfer and @xmath37 is the momentum of the @xmath1 meson .
the matrix element @xmath38 can be obtained from eq .
( [ mel2 ] ) using the identity _ = - _ ^ _ 5 .the matrix elements @xmath39 and @xmath40 can be calculated by contracting both sides of the eqs .
( [ mel1 ] ) and ( [ mel2 ] ) with @xmath41 , respectively .
we get then [ mel1q ] ( k ) [ mel2q ] |s i _
q^b b & = & i _
q^^ k^g .
using eqs .
( [ mel2])and ( [ mel2q ] ) the matrix element @xmath42 can be written in terms of form factors that are calculated in framework of the qcd sum rules @xcite as follows [ mel22q ] |s i _
q^(1+_5 ) b b & = & \ { _ ^ q^k^g_1(q^2 ) + if_1(q^2 ) } .
it should be noted that these form factors were calculated in framework of the light
front model in @xcite .
so , using eqs .
( [ mel2 ] ) , ( [ mel2q ] ) and ( [ mel22q ] ) we can easily express @xmath43 and @xmath44 in terms of the form factors @xmath45 and @xmath46 . the matrix element which describes the structure
dependent part can be obtained from eqs .
( [ mel1])([mel22q ] ) [ sd ] m_sd & = & v_tb v_tq^ * \ { |^(1-_5 ) + & + & |^(1+_5 ) + & + & i _ |^ + & + & i | _ } , where a_1 & = & ( c_br + c_sl ) g_1 + ( c_ll^tot + c_rl ) g , + a_2 & = & ( c_br - c_sl ) f_1 + ( c_ll^tot - c_rl ) f , + b_1 & = & ( c_br + c_sl ) g_1 + ( c_lr^tot + c_rr ) g , + b_2 & = & ( c_br - c_sl ) f_1 + ( c_lr^tot - c_rr ) f , + g & = & 4 c_t g_1 , + n & = & - 4 c_t ( f_1+g_1 ) , + h & = & n ( qk ) , + g_1 & = & - 8 c_te g_1 , + n_1 & = & 8 c_te ( f_1+g_1 ) , + h_1 & = & n_1(qk ) for the inner bremsstrahlung part we get [ ib ] m_ib & = & v_tb v_tq^ * e f_b i \ { f | ( - ) _ 5 + & + & f_1 | } . in obtaining this expression
we have used 0 |s _ _ 5 b b & = & - i f_b p_b , + 0 |s _ ( 1+_5 ) b b & = & 0 , and conservation of the vector current .
the functions @xmath47 and @xmath48 are defined as follows f & = & 2 m _
( c_lr^tot - c_ll^tot + c_rl - c_rr ) + ( c_lrlr - c_rllr - c_lrrl + c_rlrl ) , + f_1 & = & ( c_lrlr - c_rllr + c_lrrl - c_rlrl ) , the double differential decay width of the @xmath9 process in the rest frame of the @xmath1 meson is found to be = ^2 , where @xmath49 and @xmath50 are the photon and one of the final lepton energy , respectively .
the boundaries of @xmath49 and @xmath50 are determined from the following inequalities 0 e _
, - v e_1 + v , where v= , is the lepton velocity .
the @xmath51 term is infrared free ; interference term has an integrable infrared singularity and only @xmath52 term has infrared singularity due to the emission of soft photon . in the soft photon limit
the @xmath11 decay can not be distinguished from the pure leptonic @xmath53 decay .
for this reason , in order to obtain a finite result the @xmath54 and the pure leptonic @xmath53 decay with radiative corrections must be considered together .
it was shown explicitly in the second reference of @xcite that when both processes are considered together , all infrared singularities coming from the real photon emission and the virtual photon corrections are indeed canceled and the final result is finite . in the present work our point of view is slightly different from the standard description , namely , we consider the @xmath32 decay as a different process but not as the @xmath55 correction to the @xmath13 decay .
in other words , we consider the photon in the @xmath11 decay as a hard photon .
for this reason , in order to obtain the decay width of the @xmath56 we must impose a cut on the photon energy , which will correspond to the experimental cut imposed on the minimum energy for detectable photon .
we require the photon energy to be larger than @xmath57 , i.e. , @xmath58 , where @xmath59 .
after integrating over lepton energy , we get the following expression for the photon energy distribution [ bela ] + & & - 4 + & & - 4 + & & + m_b^2 ( 1 - 4r - x ) + & & + m_b^2 ( 1 + 8r - x ) + & & - m_b^2 ( a_1 ^2 + a_2 ^2 + b_1 ^2 + b_2 ^2 ) ( 1 - r - x ) - ( g ^2 + g_1 ^2 ) + & & + 2 m_([a_2+b_2 ] [ 6 h_1^(1-x ) + 2 g_1^x - m_b^2 n_1^x ( 1-x ) ] ) } + & & + 4 f_b \ { 2 v x ( 1-x ) + & & + ln } + & & + 4 f_b^2 \{2 v ( f ^2 + ( 1 - 4 r ) f_1 ^2 ) + & & + ln } ) , where @xmath60 is the dimensionless photon energy , @xmath61 .
it follows from eq .
( [ bela ] ) that in order to calculate the decay width explicit forms of the form factors @xmath62 and @xmath46 are needed .
these form factors are calculated in framework of light
cone @xmath63 sum rules in @xcite and @xcite , and their @xmath64 dependences , to a very good accuracy , can be represented in the following dipole forms , [ ff ] g_(q^2 ) & = & , f_(q^2 ) = , + + + g_1(q^2 ) & = & , f_1(q^2 ) = , which we will use in the numerical analysis .
in this section we will present our numerical analysis .
numerical results are presented only for the @xmath65 decay .
it is clear that in the su(3 ) limit the difference between the decay rates is attributed to the ckm matrix elements only , i.e. , ^2 .
the values of the main input parameters which have been used in the present work are : @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 . for the wilson
coefficients @xmath72 and @xmath73 we have used the results given in @xcite . in the leading logarithmic approximation , at the scale @xmath74 they are given as @xmath75 , @xmath76 .
although individual wilson coefficients at @xmath77 level are all real , the effective wilson coefficient @xmath78 has a finite phase .
the analytic expression of @xmath78 for the @xmath79 transition , in next to leading order approximation is given as [ c9 ] c_9^eff(m_b , s ) & = & c_9(m_b ) + 0.124 w(s ) + g(m_c , s ) ( 3 c_1 + c_2 + 3 c_3 + c_4 + 3 c_5 + c_6 ) + & - & g(m_q , s ) ( c_3 + 3 c_4 ) - g(m_b , s ) ( 4 c_3 + 4 c_4 + 3 c_5 + c_6 ) + & + & ( 3 c_3 + c_4 + 3 c_5 + c_6 ) , where @xmath80 , @xmath81 and the values of the individual wilson coefficients are listed in table 1 .
@xmath82 in eq .
( [ c9 ] ) @xmath83 describes one gluon corrections to the matrix element of the operator @xmath84 and the function @xmath85 stands for the one loop corrections to the four quark operators @xmath86@xmath87 with mass @xmath88 at the dilepton invariant mass @xmath89 @xcite : + & & , where @xmath90 and @xmath91 . it is well known that the wilson coefficient @xmath92 receives also long distance contributions , which have their origin in the real @xmath93 intermediate states , i.e. , @xmath94 , @xmath95 , @xmath96 ( see @xcite ) . in this work
we restrict ourselves only to short contributions .
furthermore we assume that all new wilson coefficients are real and varied in the region @xmath97 . in fig .
( 1 ) we present the dependence of the integrated branching ratio of the @xmath98 decay on the new wilson coefficients for the cut @xmath99 imposed on the photon energy , without long distance effects .
it clearly follows from this figure that as the new wilson coefficients @xmath100 and @xmath29 increase from @xmath101 to @xmath102 branching ratio decreases .
however this behavior is reversed for the coefficients @xmath103 and @xmath27 i.e. , when these coefficients increase from @xmath101 to @xmath102 branching ratio also increases accordingly .
exception to these cases takes place for the coefficient @xmath31 . in the region
@xmath104 branching ratio decreases and in the region @xmath105 it tends to increase . for the choice of the photon energy cut
@xmath106 all the previous arguments remain valid with only a slight decrease in the value of the branching ratio . from all present figures we observe that when all wilson coefficients lie in the range @xmath107 , the branching ratio is more sensitive to the existence of tensor @xmath30 , scalar @xmath108 and vector @xmath109 type interactions .
on the other side , when wilson coefficients lie in the region @xmath110 the branching ratio is more sensitive to the scalar type interaction with coefficients @xmath28 and @xmath27 .
photon energy distribution can also give useful information about new physics effects .
for this purpose , in fig .
( 2 ) we present the dependence of the differential branching ratio for the @xmath98 decay on the dimensionless variable @xmath60 at different values of tensor interaction with coefficient @xmath30 .
we observe from this figure that when @xmath111 then the related tensor interaction gives constructive contribution to the sm result , and when @xmath112 the contribution is destructive .
in other words measurement of the differential branching ratio can give essential information about the sign of new wilson coefficients .
performing measurement at different photon energies can give information not only about magnitude but also about the sign of the new wilson coefficient interaction .
note that the results presented in this work can easily be applied to the @xmath113 decay .
for example , the branching ratio for the @xmath113 decay at @xmath99 , without the long distance effects at @xmath114 is larger about 5 times , compared to that of the sm prediction of the branching ratio for the @xmath113 decay .
additionally , the dependence of the branching ratio on the new wilson coefficients is symmetric with respect to the zero point ( see fig .
it should be stressed that by studying the dalitz distribution @xmath115 at different fixed values of the final lepton ( or photon ) energies we can get useful information not only about the magnitude of the new wilson coefficients but also about their sign . in conclusion , using a general , model independent effective hamiltonian , the @xmath65 decay is studied .
it has been shown that the branching ratio and photon energy distribution are very sensitive to the existence of new physics beyond sm .
we conclude that the radiative @xmath116 decay can be measured ib the b factories as well as lhc b experiments , in which @xmath117 @xmath118 mesons are expected to be produced per year . 99 the babar physics book , eds .
p. f. harrison and h. r. quinn , + slac report ( 1998 ) 504 .
belle collaboration , e. prebys _ at al _ , _ nucl .
_ * a446 * ( 2000 ) 89 .
g. burdman , t. goldman and d. wyler , _ phys .
_ * d51 * ( 1995 ) 111 .
g. eilam , i. halperin and r. r. mendel _ phys .
lett . _ * b361 * ( 1995 ) 137 .
p. colangelo , f. de fazio , g. nardulli , _ phys .
_ * b372 * ( 1996 ) 311 .
d. atwood , g. eilam , a. soni , _ mod .
lett . _ * a11 * ( 1996 ) 1061 .
t. m. aliev , a. zpineci and m. savci , _ phys .
* b393 * ( 1997 ) 393 . c. q. geng , c. c. lih and w. m. zhang , _ phys .
* d57 * ( 1998 ) 5697 .
g. p. korchemsky , d. pirjol and tung mow yan , _ phys .
* d61 * ( 2000 ) 114510 . t. m. aliev , a. zpineci and m. savci , _ phys . rev . _
* d55 * ( 1997 ) 7059 ; + t. m. aliev , n. k. pak and m. savci , _ phys .
lett . _ * b424 * ( 1998 ) 175 .
g. eilam , c. d. l , d. x. zhang , _ phys .
* b391 * ( 1997 ) 461 .
t. m. aliev , c. s. kim , y. g. kim , _ phys .
_ * d62 * ( 2000 ) 014026 ; + t. m. aliev , d. a. demir , m. savci , _ phys . rev . _ * d62 * ( 2000 ) 074016 ; + s. fukae , c. s. kim , t. morozumi , t. yoshikawa , _ phys . rev . _ * d59 * ( 1999 ) 074013 .
c. q. geng , c. c. lih and wei min zhang , _ phys .
_ * d62 * ( 2000 ) 074017 g. buchalla , a. j. buras and m. e. lautenbacher , _ rev .
* 68 * ( 1996 ) 1125 ; a. j. buras and m. mnz , _ phys .
_ * d52 * ( 1995 ) 186 .
m. misiak , _ nucl
. phys . _ * b398 * ( 1993 ) 23 ; erratum : _ ibid .
_ * b439 * ( 1995 ) 461 . c. s. lim , t. morozumi and a. i. sanda , _ phys .
* b218 * ( 1989 ) 343 ; + n. g. deshpande , j. trampetic and k. panose , _ phys . rev .
_ * d39 * ( 1989 ) 1461 ; + a. i. vainshtein , v. i. zakharov , l. b. okun and m. a. shifman , + _ sov .
* 24 * ( 1976 ) 427 .
* fig . ( 1 ) * the dependence of the integrated branching ratio of the @xmath119 decay on the new wilson coefficients for the cut @xmath99 imposed on the photon energy , only for short distance effects .
+ + * fig .
( 2 ) * the dependence of the differential branching ratio for the @xmath119 decay on the dimensionless variable @xmath60 at different values of tensor interaction with coefficient @xmath30 , without the long distance effects .
+ + * fig .
( 3 ) * the same as in fig . ( 1 ) , but for the @xmath120 decay . | using the most general , model independent form of the effective hamiltonian rare decays @xmath0 are studied .
the sensitivity of the photon energy distribution and branching ratio to the new wilson coefficients is investigated .
16.3 true cm 23.0 true cm -0.8 true in 0.00 true in pacs number(s ) : 12.60.i , 13.20.v , 13.20.he |
in this article we apply a method of quantum many - body theory called the coupled cluster method ( ccm ) @xcite to study strongly interacting quantum spin systems .
the ccm is not restricted , in principle , by the spatial dimensionality of the problem or by the presence of competition between bonds , i.e. , in frustrated quantum spin systems .
an important advance in the accuracy of the method for a localized approximation scheme called the lsub@xmath1 scheme has been afforded by the use of `` high - order '' ccm via computer - algebraic implementations @xcite .
this computer code developed by djj farnell and j schulenburg @xcite is very flexible in terms of the range of underlying crystallographic lattices , values for spin quantum number , and types of hamiltonian that may be studied .
a common first task in the practical application of the ccm is to rotate the local spin axes of the ( often classical ) model state so that ( notationally only ) the spins all appear to point in the downwards @xmath0-direction .
although the hamiltonian is changed by transforming these local spin axes , these rotations are unitary and so they do not affect the energy eigenvalues or expectation values .
furthermore , we note that hitherto only coplanar model states have been used because they do not lead to complex numbers in the new hamiltonian . by contrast
, three dimensional ( 3d ) non - coplanar model states can lead to imaginary terms in the new hamiltonian after rotation and so are more difficult to treat computationally . in principle , however , macroscopic quantities predicted by the ccm ( such as the ground - state energy and order parameter ) should be still real even though the hamiltonian is now complex because again these transformations of local spin axes are unitary . here
we explain how we may use such 3d model states for the ccm and how we may solve for the ( now possibly complex ) ccm correlation coefficients . as an illustration of the method
, we consider the spin - half triangular - lattice heisenberg antiferromagnet in a magnetic field @xcite .
as is well - known , the response of quantum magnetic systems to an external field is revealed by its magnetization curve .
the magnetic processes of quantum anitferromagnets is discussed , e.g. , in refs .
@xcite , and the interested reader is referred to these sources for more details .
the hamiltonian that we will use here is given by @xmath2 where the index @xmath3 runs over all lattice sites on the triangular lattice .
the expression @xmath4 indicates a sum over all nearest - neighbor pairs , although each pair is counted once and once only .
the strength of the applied external magnetic field is given by @xmath5 .
there are two ground states classically ( shown in fig .
[ model_states ] ) : a set of coplanar states and a single non - coplanar state .
the quantum system is discussed in refs .
@xcite . although on the classical level both cases ( coplanar and non - coplanar ) are energetically equivalent@xcite , previous results of approximate methods indicate that thermal or quantum fluctuations ought to favor the planar configuration @xcite .
previous results of the ccm @xcite indicate that a plateau state occurs for @xmath6 .
( note that we compare new results presented in this article for the non - coplanar states to those earlier results of ref .
these results for the plateau are in excellent agreement with experimental results for the magnetic compound ba@xmath7cosb@xmath8o@xmath9 ( a spin - half triangular - lattice antiferromagnet ) , which demonstrates a spin plateau that agrees quantitatively with results of the ccm and exact diagonalizations @xcite .
furthermore , ccm results indicate that a similar plateau occurs over the range @xmath10 for the for the spin - one triangular - lattice antiferromagnet , and this theoretical result has subsequently been established experiment for the compound ba@xmath7nisb@xmath8o@xmath9 ( a spin - one triangular - lattice antiferromagnet ) @xcite . the main goal of our paper is to explain how the ccm can be used with 3d model states .
firstly we present a brief description of the ccm formalism and its application via computational methods to the subject of quantum spin models with 3d model states . as an illustration of our method
, we describe the application of the method to the spin - half heisenberg model for the triangular lattice at zero temperature in the presence of an external magnetic field .
we present our results and then discuss the conclusions of this research .
as the ccm has been discussed extensively elsewhere ( see refs .
@xcite ) , a brief overview of the method is presented here only . note however that the solution of the ccm equations for the case of 3d model states is presented in an appendix , which has not been attempted before . in this case , ccm correlation coefficients may be complex , and so extensive changes to the basic computer code that implements the ccm to high orders of approximation are necessary , again as described in the appendix .
we begin the brief overview of the ccm method by presenting the ground - state schrdinger equations , which are given by @xmath11 and bra and ket states are given by @xmath12 we use model states ( denoted @xmath13 in the ket state and @xmath14 in the bra state ) as references states for the ccm and those used here are shown in fig .
[ model_states ] .
we note that the ccm ket- and bra - state equations are given by @xmath15 { \rm e}^s|\phi\rangle & = & 0 , \;\ ; \forall i \neq 0 \;\ ; . \label{bra_state_eqn}\end{aligned}\ ] ] and that the method in which eqs .
( [ ket_state_eqn ] ) and ( [ bra_state_eqn ] ) are solved has been discussed extensively elsewhere @xcite and so is not given here .
the ground - state energy is given by @xmath16 this equation is a function of the ket - state correlation coefficients only .
we differentiate between those the model states that are coplanar and the single model state that is non - coplanar ( or `` 3d '' ) .
the analysis for the coplanar model state is given in ref .
@xcite and we refer the interested reader to this publication for more details .
the non - coplanar model state iv has spins that make an angle @xmath17 to the plane perpendicular to the external field , as is also shown in fig .
[ model_states ] .
we rotate the local spin axes of the spins such that all spins appear to point downwards for all four model states i , ii , iii , and iv in fig .
[ model_states ] .
we have four new hamiltonians after rotation of the local spin axes of the spins such that all spins appear to point downwards for all four model states i , ii , iii , and iv in fig .
[ model_states](a - d ) .
the hamiltonian for model state i , fig .
[ model_states ] after rotation of the local spin axes is given by : @xmath18 where the sum @xmath19 goes from sublattice @xmath20 to sublattice @xmath21 ( and with directionality ) .
note that @xmath22 indicates a sum that goes from sublattice @xmath21 to sublattice @xmath23 and sublattice @xmath23 to sublattice @xmath20 , respectively ( and with directionality ) .
the hamiltonian for model state iii , fig .
[ model_states ] after rotation of the local spin axes is given by : @xmath24 where the sum @xmath25 goes from sublattice @xmath23 to sublattices @xmath20 and @xmath21 ( with directionality ) and @xmath26 goes over each bond connecting the @xmath20 and @xmath21 sublattices , but counting each one once only ( and without directionality ) .
we note that we have three sites in the unit cell for all of the models states used for the triangular lattice antiferromagnet .
the hamiltonian for model state ii , fig .
[ model_states ] , is a limiting case of eqs .
( [ roth2 ] ) and ( [ roth3 ] ) .
the hamiltonian for the non - coplanar model state iv after rotation of the local spin axes is given by : @xmath27 where @xmath28 are those `` directional '' nearest - neighbor bonds on the triangular going from the @xmath20 sublattice to @xmath21 sublattice , @xmath21 sublattice to @xmath23 sublattice , and @xmath23 sublattice to @xmath20 sublattice ( in those directions only and not reversed ) .
we see that this hamiltonian now contains `` real and imaginary components '' .
henceforth , we shall take the expression `` real and imaginary components '' to mean that the rotated hamiltonian contains explicit factors involving the imaginary number @xmath29 and other explicit factors that do not involve this imaginary number .
we note that @xmath17 is the angle that the spins make to the plane perpendicular to the applied external field .
full details of the rotations used in the derivation of this hamiltonian are presented in appendix b. the manner in which ccm equations may be solved for the bra- and ket - state equations when correlation coefficients are allowed to be complex is discussed in the appendix , although we note that the problem essentially reduces to a doubling of the number of ccm equations to be solved ( i.e. , for the real and imaginary components separately ) .
however , we note that we use the lsub@xmath1 approximation ( in which clusters of @xmath1 contiguous sites limited to @xmath1 spin flips are included in @xmath30 and @xmath31 ) and that we consider the angles as free parameters in the ccm calculation . these angles are found by direct minimization of the ccm ground - state energy .
this was achieved computationally at a given level of lsub@xmath1 approximation , and a minimum ground state energy with respect to these canting angles was also found computationally for a given fixed value of @xmath5 .
we note that there was only one angle was needed for model states i and iv , whereas two such angles were needed for model state iii .
values of @xmath5 were varied incrementally and the minimization process of the energy with respect to the canting angles repeated .
the ccm calculations are costly in terms of computing time because we needed to minimize the ground - state energy with respect to such angles at each value of @xmath5 .
we remark that the we have twice as many equations to solve for the 3d model state ( for the real and imaginary components separately ) `` as normal '' for the ccm . for this reason ,
results up to the lsub6 level of approximation only are quoted in these initial tests , although we find that ground - state energies are highly converged for all model states even at this relatively low level of approximation . in order to investigate the magnetization process in antiferromagnets
, we consider the total lattice magnetization @xmath32 along the direction of the magnetic field .
we note that the important hellmann - feynman theorem is obeyed by the ccm and so we may obtain the lattice magnetization by finding @xmath33 , which is carried out computationally in this paper .
the method by which other expected values may be found is also discussed in the appendix .
as mentioned above , we present ccm results for model states i , ii , iii , and iv shown in fig . [ model_states ] .
the computational effort of the ccm calculations presented here for the non - coplanar model state iv to very high orders is very great and so results up to lsub6 are presented here only for this model state in these initial studies .
results of the coplanar model states i to iii from ref .
@xcite up to lsub6 are also presented here for the purposes of comparison only , although we note that higher orders of approximation than lsub6 were carried out in ref .
@xcite for these states .
the results for the ground - state energy are shown in fig .
[ triangle_energies ] .
we note that the results for the coplanar model states i to iii from ref .
@xcite with lowest energy are shown only as a function of @xmath5 in fig .
[ triangle_energies ] .
thus , results of model state i only are presented for small values of the applied magnetic field strength @xmath5 and results of model state iii only are presented for higher values of @xmath5 near to @xmath34 .
the results of both model states coincide in the intermediate regime .
again , these lsub@xmath1 series of results are found to converge rapidly with increasingly levels of lsub@xmath1 approximation over all values of the external field parameter @xmath5 .
results for model state iv are also shown .
firstly , we note that the imaginary component of the ground - state energy is found to sum to zero , as expected and required .
furthermore , the results for the ground - state energies for model state iv are much higher in value than their coplanar counterparts at identical levels of lsub@xmath1 approximation .
note that lsub6 was the highest level of approximation possible for this model state in these initial tests and so we limit all presented results ( for all model states ) to this level of approximation in order to allow a direct and unbiased comparison .
however , we see that results for the ground - state energy are highly converged even at the lsub6 level of approximation ( by comparing results of lsub6 to lsub5 and lsub4 in fig .
[ triangle_energies ] ) and so these results are adequate to establish that the ground - state energies are indeed lower for the coplanar case .
as noticed previously @xcite , these results provide clear and strong evidence that the ground state of this system is coplanar .
this case has therefore been an excellent first test of new ccm based on a non - coplanar `` three - dimensional '' model state .
the results for the total lattice magnetization for model states i to iii from ref .
@xcite are shown in fig .
[ triangle_magnetization ] .
the lsub@xmath1 results are again seen to converge rapidly for increasing @xmath1 .
however , there is a radical departure from the classical straight - line behavior ( i.e. @xmath35 ) in this case .
these previous ccm results for the coplanar states accurately detect the plateau in the @xmath32 versus @xmath5 curve at @xmath36 . indeed
, these previous results of the ccm @xcite for the coplanar model states carried out to the lsub8 level of approximation indicate that the width of this plateau is given by @xmath6 .
( note that the plateau corresponds to the straight " part of the curve in the @xmath37 curve shown in fig .
[ triangle_energies ] . ) by contrast , we note that the new results from model state iv show no such spin plateau any level of lsub@xmath1 attempted ( e.g. , lsub4 , lsub5 and lsub6 shown in the fig . [ triangle_magnetization ] ) .
note that this plateau has been observed for this system by other approximate methods and in experiment @xcite .
we note importantly again that the lattice magnetization must be real - valued because we find values for this quantity by taking first derivative of the ground - state energy , which itself is found to be real - valued ( and not complex ) for all values of @xmath5 .
all of these results are excellent corroborating evidence that the ground state is not of the type shown by model state iv ( i.e. , non - coplanar ) , but is rather of coplanar type shown by model states i to iii . again , this is an excellent first check of the method for 3d model states .
in this article we described how the coupled cluster method ( ccm ) may be applied to study the behavior of quantum magnets with the aid of non - coplanar `` 3d '' model states .
a ( slightly ) modified method for solving for the ccm equation by `` direct iteration '' and for complex ccm correlation coefficients is presented in an appendix .
we employed new computer code to find the ground - state energy and the total lattice magnetization of the spin - half triangular - lattice antiferromagnetic systems in the presence of external magnetic fields for both coplanar and non - coplanar model states ( which are the classical ground states ) .
it was found that the ground - state energy was real - valued for the non - coplanar model state , as expected and required .
results up to lsub6 were possible only in these initial tests for the non - coplanar model state due to the increased computational complexity of the problem .
however , these results were clearly highly converged even at this level of approximation . in agreement with previous results of other methods ,
the coplanar states were found to have lower energy .
furthermore , the spin plateau for @xmath36 known to occur in this system was detected by the coplanar model states , although it was not detected by the non - coplanar model state at any level of approximation .
we conclude that the spin - half triangular lattice antiferromagnet in an external magnetic field for the non - coplanar model state has been an excellent first test of new ccm based for a non - coplanar `` three - dimensional '' model state .
the method by which 3d model states may be employed for the high - order ccm code is discussed below .
the results for the application of this new approach to the spin - half triangular lattice antiferromagnet in an external magnetic field are presented in the main text .
normally we solve the ccm equations either by using newton - raphson method or by directly iterating the ccm equations .
the direct iteration method for the ket - state is achieved by casting the @xmath38 basic ccm equations ( each corresponding to one of the @xmath38 fundamental clusters at a given level of approximation ) , namely : @xmath39 which we write as @xmath40
we then rearrange this equation into the form @xmath41 hence , given an initial starting solution for the set of correlation coefficients @xmath42 , we iterate the following equation for all of the @xmath38 equations to convergence : @xmath43 this simple approach is successful in solving the ccm equations when all of the correlation coefficients are real .
however , for 3d model states we can not ensure that the hamiltonian is real - valued after rotation of the local spin axes .
indeed , in most cases it will not be so .
the extension of the direct - iteration approach to complex solutions for the ccm ket - state correlation coefficients might be achieved computationally by using complex data type ( and all operators ) using c++ classes .
however , the changes to the ccm code would be extensive .
a simpler alternative is just to break the @xmath38 ccm equations to @xmath44 equations , corresponding to the real and complex parts of the equation(s ) governed by eq .
( [ equation1 ] ) .
hence , we split @xmath45 into real and imaginary components , i.e. , @xmath46 , as well as @xmath47 and @xmath48 such that @xmath49 we now iterate these ( now ) @xmath44 equations for @xmath50 and @xmath51 ( which correspond to the real and imaginary parts of ket - state correlation coefficient @xmath45 ) to convergence .
note that we use the following products identities ( up to fourth order ) in the ccm equations of eq .
( [ directitimag ] ) , namely : @xmath52 these identities have been coded directly into the high - order ccm code in order to solve for complex ket - state coefficients ; although this is somewhat tedious , it has the advantage of being more straightforward and probably a little speedier at runtime than using complex number classes to solve this problem .
ground - state energies are found by evaluating @xmath16 we note that the ground - state energy @xmath53 is a function of the ccm coefficients up to quadratic terms in these coefficients ( i.e. , terms such as const . , @xmath54 and @xmath55 ) .
hence , we use the above equations in order to find the real and imaginary parts of such terms . furthermore , as the hamiltonian has both real and imaginary parts we divide the ground - state energies via : @xmath56 after determining the real and imaginary components of the ccm ket - state correlation coefficient , we ought to find that @xmath57 , as all of the terms contributing to it should cancel each other out .
this is because that are real before rotation of local spin axes ( and must @xmath53 clearly is real number ) must still be real even though we are now solving for complex ccm correlation coefficients .
indeed , this constraint , i.e. , @xmath57 , forms an excellent check that our computer code is working correctly .
this was found to be the case for model state iv employed here for the triangular - lattice heisenberg antiferromagnet for all values of the external magnetic field strength , @xmath5 .
we determine the bra - state correlation coefficient firstly by finding @xmath58 and then by taking the partial derivative of this expression with respect to @xmath45 .
we note again that the bra - state operator is defined via @xmath59 and so we see that we obtain a linear equation @xmath60 however , we note again that @xmath61 and @xmath62 and so we may write the above bra - state equation as @xmath63 this is just an equation that is linear in terms of the bra - state correlation coefficients . if we now write @xmath64 then this equation is solving readily by putting the @xmath65 terms for the @xmath66 bra - state equation on the left - hand side of the equation and everything else on the right - hand side .
we then iterate to convergence once again .
previously , bra - state correlation coefficients were real - valued , although we can not preclude that the bra - state coefficients are complex for 3d model states .
once again we solve this problem by iterating @xmath44 bra - state equations , corresponding to the real and imaginary parts of the bra - state correlation coefficients . in this case
, we have terms such as : @xmath67 as the well - known helmann - feynman theorem is obeyed by the ccm , the lattice magnetization is found by taking the first derivative of the ground - state energy per spin with respect to the external field strength @xmath5 .
this was achieved computationally in this article as this approach is simple and straightforward .
an alternative approach ( not carried out here ) may also be used to find the lattice magnetization @xmath68 .
we begin by rotating the local coordinates of the spin axes and we form the relevant operator @xmath68 , which may well now be complex . however , once we have solved for both the ket - and bra - state equations we should find that the imaginary part of this quantity should sum to zero , as was indeed seen for the ground - state energy .
we note that the local `` sublattice '' magnetization @xmath69 ( where @xmath70 is expressed in spin coordinates _ after _ rotation of the local spin axes ) is found to be complex .
this seems perplexing at first because we are finding @xmath69 , i.e. , there appears to be no imaginary term inside this expectation value ! however , we see that @xmath70 is expressed in terms of the _ rotated _ spin axes . if we were to carry out the _ inverse _ transformation ( i.e. , from the rotated spin axes to the original axes ) then @xmath70 would , of course , probably become complex - valued when expressed in terms of these _ original _ spin axes .
we might well now expect this new quantity to be complex - valued , and indeed this was found to be the case .
caution should therefore be exercised in attaching importance to the `` sublattice magnetization '' for such 3d model states , although it might be well - defined in certain circumstances .
this is quite different to coplanar 2d model states for the ccm , and for which the sublattice magnetization is always real . to summarize , however , only those macroscopic observables that are real - valued in the original spin coordinates for 3d model states ought definitely to be real - valued once we have solved for the ket- and bra - state equations and formed the relevant expectation values . in these cases , imaginary components present after rotation of local spin axes should sum to zero when evaluating the expectation value .
we now present a derivation of the hamiltonian after rotation of the local spin axes .
we begin by again presenting the hamiltonian before rotation of the spin axes : @xmath71 where the index @xmath3 runs over all lattice sites on the triangular lattice .
the expression @xmath4 indicates a sum over all nearest - neighbor pairs , although each pair is counted once and once only .
the strength of the applied external magnetic field is given by @xmath5 .
note that we assume that the external field applies now in the @xmath72-direction rather than the @xmath0-direction .
this is a notational change only in order to simplify somewhat the following mathematics and it does not affect our results .
indeed , this choice allows us to use the model state with zero field now lies in the @xmath73-plane , as used in ref .
@xcite . the classical ground - state of eq .
( [ newh ] ) when @xmath74 is the nel - like state where all spins on each sublattice are separately aligned ( again as in ref .
@xcite , all in the @xmath73-plane , say ) . the spins on sublattice a are oriented along the negative _ z_-axis , and spins on sublattices b and c are oriented at @xmath75 and @xmath76 , respectively , with respect to the spins on sublattice a. we perform the following spin - rotation transformations .
specifically , we leave the spin axes on sublattice a unchanged , and we rotate about the @xmath72-axis the spin axes on sublattices b and c by @xmath76 and @xmath75 respectively , @xmath77 this leads to a hamiltonian given by : @xmath78 we note that the summation in eq .
( [ hstage1 ] ) again runs over nearest - neighbor bonds , but now also with a _
directionality _ indicated by @xmath79 , which goes from a to b , b to c , and c to a. spins now appears mathematically to point in the downwards @xmath0-direction when the external field is zero ( @xmath74 ) . for @xmath80
, the spins appear now to be at an angle of @xmath17 to the negative @xmath0-axis in the @xmath81-plane .
thus , we rotate the local spin axes additionally by an amount @xmath82 about the @xmath83-axis by using the following ( passive ) transformation : @xmath84 this transformation yields a hamiltonian of : @xmath85 all spins now appear to the lie in the negative @xmath0-direction for any given value thus of @xmath17 ( and thus also for all values of the external field ) .
we now use the following expressions : @xmath86 such that the final version of the hamiltonian is obtained : @xmath87 the model state @xmath88 now consists of spins that appear ( mathematically only ) to lie along the negative @xmath0-axis for all values of @xmath17 .
we note that by using @xmath89 and also by using the expressions @xmath90 we obtain the classical ground - state energy by using eq .
( [ finalh ] ) , namely , of @xmath91 , as required .
this is an excellent initial check of the the hamiltonian of eq .
( [ finalh ] ) after all of the rotations of the local spin axes have been carried out .
m. roger and j.h .
hetherington , phys .
b * 41 * , 200 ( 1990 ) ; _ ibid _ europhys . lett . *
11 * , 255 ( 1990 ) .
r.f . bishop , j.b .
parkinson , and y. xian , phys .
b * 44 * , 9425 ( 1991 ) .
y. xian , j. phys . : condens .
matter * 6 * , 5965 ( 1994 ) . c. zeng , d.j.j .
farnell , and r.f .
bishop , j. stat .
* 90 * , 327 ( 1998 ) .
r.f . bishop , d.j.j .
farnell , s.e .
krger , j.b .
parkinson , j. richter , and c. zeng , j. phys . : condens .
matter * 12 * , 7601 ( 2000 ) .
farnell , k.a .
gernoth , and r.f .
bishop , j. stat .
phys . * 108 * , 401 ( 2002 ) .
farnell , j. schulenberg , j. richter , and k.a .
gernoth , phys .
b * 72 * , 172408 ( 2005 ) .
s. krger , r. darradi , j. richter , and d.j.j .
farnell , phys .
b * 73 * , 094404 ( 2006 ) .
r. darradi , o. derzhko , r. zinke , j. schulenburg , s.e .
krger , and j. richter , phys .
b * 78 * , 214415 ( 2008 ) d.j.j .
farnell , j. richter , r. zinke , and r.f .
bishop , j. stat . phys . * 135 * , 175 ( 2009 ) .
h. nishimori and s. miyashita , j. phys .
japan * 55 * , 4448 ( 1986 ) .
a.v . chubukov and d.i .
golosov , j. phys . : condens .
matter * 3 * , 69 ( 1991 ) .
j. alicea , a.v .
chubukov , o.a .
starykh , phys .
102 , 137201 ( 2009 ) .
m. oshikawa , m. yamanaka , and i. affleck , phys .
* 78 * , 1984 ( 1997 ) .
j. schulenburg and j. richter , phys . rev .
* 65 * , 054420 ( 2002 ) .
j. schulenburg , a. honecker , j. schnack , j. richter , and h .- j .
schmidt , phys .
lett . * 88 * , 167207 ( 2002 ) ; j. richter , j. schulenburg , a. honecker , j. schnack , and h.j .
schmidt , j. phys . : condens .
matter * 16 * , s779 ( 2004 ) .
t. ono , h. tanaka , h. aruga katori , f. ishikawa , h. mitamura , and t. goto , phys .
b * 67 * , 104431 ( 2003 ) .
cabra , m.d .
grynberg , p.c.w .
holdsworth , a. honecker , p. pujol , j. richter , d. schmalfu , and j. schulenburg , phys .
b * 71 * , 144420 ( 2005 ) .
r. schnalle and j. schnack , phys .
b * 79 * , 104419 ( 2009 ) . c. schrder , h. nojiri , j. schnack .
p. hage , m. luban , and p. kgerler , phys .
* 94 * , 017205 ( 2005 ) . | the coupled cluster method ( ccm ) has previously been applied to study the ground- and excited - state properties of many different types of frustrated and unfrustrated quantum spin systems .
a common feature in the application of the ccm is to rotate the local spin axes of the ( often classical ) model state so that ( notationally only ) the spins all appear to point in the downwards @xmath0-direction .
hitherto , we remark that only coplanar model states have been used because they do not lead to imaginary terms in the new hamiltonian . by contrast , non - coplanar `` three - dimensional '' ( 3d ) model states can lead to imaginary terms in the new hamiltonian after rotation . in principle , however , macroscopic quantities predicted by the ccm ( such as the ground - state energy and order parameter ) should still be real ( even though the hamiltonian may be complex ) because the transformations of local spin axes are unitary . here
we explain how we may use such 3d model states for the ccm and how we may solve for the ( now possibly complex ) ccm correlation coefficients .
we present results for the spin - half triangular - lattice antiferromagnet in an external magnetic field .
we use both coplanar model states and a single 3d non - coplanar model state .
we calculate the ground - state energy and the lattice magnetization as a function of the magnetic field strength .
we find that the energies of the coplanar states lie lower than that of the non - coplanar state for all values of the external field , as seen in other approximate studies for this model .
we find that the non - coplanar state does not detect the well - known spin plateau occurring in this model , although ( as seen before ) it is clearly observed for the coplanar model states using the ccm .
these results are an excellent initial validation of the new approach for the application of the ccm using 3d model states . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.